The following abstracts are for PhD and Master's theses written using global NCAR 
climate models since the release of CSM1 in 1996.  The list is current through June 2005.

Title: Quantitative paleoclimatic reconstructions for the Late
           Quaternary of southern South America based on calibration
           of modern pollen and climate relationships
   Pub No: 9827679
   Author: Anderson, Lysanna
   Degree: PhD
     Date: 1998
    Pages: 219
  Adviser: Markgraf, Vera
     ISBN: 0-591-80013-6
   Source: DAI-B 59/03, p. 1017, Sep 1998
  Subject: GEOLOGY (0372); PALEOECOLOGY (0426); PALYNOLOGY (0427)
 Abstract: Three approaches to deriving quantitative estimates of
           past temperature and precipitation (multiple regression,
           response surfaces, and the method of modern analog) were
           evaluated using a modern pollen and climate data set from
           southern South America. The results of this analysis show
           that, under conditions of sparse climate data, response
           surfaces are the most robust approach. The three
           calibrated data sets were then applied to four pollen
           records, covering the last glacial - interglacial
           transition, to derive quantitative estimates of changes
           in temperature and precipitation for the South American
           mid-latitudes (41$/sp/circ$). Although there are
           variations among the records, there is generally good
           agreement, indicating largely consistent regional scale
           change. Comparisons of the quantitative climate
           reconstructions with qualitative records of
           paleoenvironmental conditions in the region show good
           Data from the Lake District of Chile and Argentina were
           then integrated with quantitative paleotemperature
           records from the equatorial region of South America, and
           with temperature estimates from the Byrd (Antarctica)
           $/delta/sp[18]$O record. This paleoenvironmental data set
           was compared to general circulation model (GCM) scenarios
           of climate change. Three different sets of model output
           from the GIST-NESS GCM for the LGM, each with different a
           different set of sea surface temperature (SST) values
           were compared. The model run using calculated values for
           SST based on an ocean heat convergence (OHC) and ocean
           heat transport (OHT) model reproduced the most accurate
           surface temperature values for the equatorial and polar
           regions. However, this model did not perform well in the
           mid-latitudes $(30/sp/circ$-60$/sp/circ$S). Model output
           from the NCAR community climate model (CCM1) for time
           slices spanning from the LGM to 6000 years ago were then
           compared to the paleoenvironmental data set. This
           comparison shows that the model's linear response to
           changing boundary conditions is not an accurate
           reconstruction of the pattern of climate change that has
           occurred since the LGM. The paleoenvironmental data show
           that the onset of warming occurred first at the high and
           mid-latitudes, preceding equatorial warming by as much as
           7000 years. The paleoenvironmental data also show that
           the continuous warming shown in the model output is
           inaccurate. Warming occurred abruptly, and reversals are
           recorded at all latitudes. In the polar region,
           temperatures jumped 5$/sp/circ$C from 21 Ka to 16 Ka, and
           had reached modern conditions by 14 Ka. At the mid-
           latitudes, temperatures transitioned from the full
           glacial minimum to warmer than today between 21 Ka and 16
           Ka, while at the equator, temperatures did not start to
           warm until ca. 14 Ka.
    Title: Role of atmosphere-ocean interaction in the midlatitude
           North Atlantic on interannual climate variability
   Pub No: 9611415
   Author: Bhatt, Uma Suren
   Degree: PhD
     Date: 1996
    Pages: 174
  Adviser: Houghton, David D.; Battisti, David S.
   Source: DAI-B 57/01, p. 397, Jul 1996
 Abstract: The primary mode of observed interannual variability in
           the North Atlantic is characterized by a north-south
           oriented dipole pattern in anomalies of surface air
           temperature and sea surface temperature (SST). The
           atmospheric circulation associated with the dipole mode
           of variability is consistent with the notion of the
           atmosphere forcing the ocean. The impact of air-sea
           interaction on the dipole-like mode of variability is
           examined using a mixed layer model (MLM) of the upper
           ocean in the North Atlantic between 20-60$/sp/circ$N
           coupled to the NCAR Community Climate Model (CCM1). The
           climatology of the MLM ocean temperature is adjusted to
           be consistent with the SSTs that form the lower boundary
           conditions for CCM1 by including heat flux corrections in
           the net forcing of the ocean. Heat and salt flux
           corrections are calculated in a series of uncoupled
           simulations where the MLM is forced with CCM1 surface
           The natural variability in a 31-year integration of the
           MLM in the North Atlantic coupled to CCM1 is compared to
           a CCM1 control simulation of similar length with SSTs
           specified to have the same climatological annual cycle as
           in the coupled integration. The mean December to February
           (DJF) climatology is essentially unchanged with the
           inclusion of midlatitude air-sea interaction. However,
           air-sea interaction leads to significant increases in the
           persistence of air temperature anomalies on interannual
           as well as monthly time scales.
           In the model subpolar North Atlantic, air and ocean
           temperature anomalies are significantly autocorrelated
           (0.4 to 0.6) from one winter to the following winter.
           These autocorrelations are consistent with the 'Re-
           emergence' mechanism (Namias and Born, 1970). Deep ocean
           temperature anomalies, present at the end of one winter,
           remained sequestered below the shallow summer mixed
           layer. As the mixed layer deepens during the following
           fall, ocean temperature anomalies from the previous
           winter are reincorporated into the surface layer.
           February mixed layer ocean temperatures are strongly
           correlated with submixed layer ocean temperatures during
           the following summer and surface ocean temperature
           anomalies the following winter. An uncoupled sensitivity
           experiment is performed, in which the MLM is forced with
           heat, momentum, and freshwater fluxes from the coupled
           simulation, and anomalies are supressed in mixed layer
           depth and entrainment heating. The sensitivity experiment
           finds that the autocorrelations from one winter to the
           next in ocean temperature are weak when anomalies in
           entrainment heating or mixed layer depths are supressed.
           The 'Re-emergence' mechanism does not play an important
           role in the subtropical model domain. Observed air and
           ocean temperature autocorrelations from one winter to the
           next are strong (weak) in the northern (southern) part of
           the North Atlantic basin, in agreement with model
           Air temperature anomalies decay more slowly as a result
           of air-sea interaction. Short-term autocorrelations of
           air temperature are statistically significant up to 1-2
           and 3-4 months in the control and coupled simulations,
           respectively. When air temperatures are not affected by
           SST anomalies as in the control simulation, the turbulent
           heat fluxes act to strongly damp air temperature
           anomalies (Frankignoul, 1985). With the inclusion of
           variable ocean temperatures, the 'thermal damping' of
           atmospheric temperature anomalies is weaker. Consistent
           with these findings, the variance of air temperature
           increases and of total heat flux decreases as a result of
           coupling. (Abstract shortened by UMI.)
    Title: Global optimization of numerical models by simulated
           annealing (Weather prediction)
   Pub No: 9957125
   Author: Campbell, William Francis
   Degree: PhD
     Date: 1999
    Pages: 212
  Adviser: Baer, Ferdinand
     ISBN: 0-599-60324-0
   Source: DAI-B 60/12, p. 6156, Jun 2000
 Abstract: The problems of vertical level placement and three-
           dimensional truncation of numerical models are
           investigated. Because the atmosphere is inhomogeneous,
           anisotropic, and unbounded above, there are no natural
           basis functions in the vertical upon which to base
           spectral methods for numerical prediction. A finite
           difference formulation in the vertical, with levels
           typically chosen to match observational levels, is used
           by most forecast and climate models. Our method computes
           the number of vertical levels as a function of horizontal
           scaling, and selects their location as a function of
           vertical coordinate system and atmospheric data.
           The central premise of our research is that the optimal
           grid for computation is the grid that yields the best
           linear spline fit to the meteorological fields of
           interest. A continuous “ground truth” for
           each field was constructed from analyzed wind and
           temperature data, via cubic splines. The best linear
           spline fit was then found by minimizing the integral of
           the squared difference between arbitrary linear splines
           and ground truth, via simulated annealing. Using NCAR's
           Community Climate Model (CCM3), grids produced with
           seasonal mean data for summer and winter were tested
           against each other and against the standard grid at the
           same average resolution.
           Allowing resolution to change dynamically between the
           horizontal and the vertical enables the direct prediction
           of the optimum truncation ratio. Because the energy of
           the grid is affected only by the <italic>distribution
           </italic> of curvature in each dimension, simulated
           annealing, meteorological data, and the linear spline
           hypothesis suffice to solve the level placement and
           relative truncation problems with no additional
           The applicability of this procedure is quite general:
           given the characteristic structures of any set of PDE's,
           the optimal predictive grid can be constructed. In
           addition, a global optimization method that is heretofore
           absent from the meteorological literature, but of general
           utility, is employed. Most importantly, a specific,
           testable prediction of truncation ratio is produced.
    Title: Climate variability in the North Atlantic on decadal and
           multi-decadal time scales:  A numerical study
   Pub No: 3001162
   Author: Cheng, Wei
   Degree: PhD
     Date: 2000
    Pages: 154
  Adviser: Bleck, Ranier; Rooth, Claes G. H.
     ISBN: 0-493-09584-5
   Source: DAI-B 62/01, p. 133, Jul 2001
           SCIENCE (0608)
 Abstract: The goal of this work is to understand the mechanisms
           that drive the decadal and multi-decadal climate
           variability in the North Atlantic. Natural climate
           variability on these particular time scales occupies a
           central position in discussions of anthropogenic climate
           changes, but many aspects related, to this issue are
           still poorly understood.
           The major tool used in this study is a coupled general
           circulation model consisting of the NCAR CCM3 and the
           Miami Isopycnic Coordinate Ocean Model. The simulated
           decadal variability in the North Atlantic is dominated by
           a tri-pole pattern in sea surface temperature and a North
           Atlantic Oscillation (NAO) pattern in sea level pressure.
           The associated oceanic fluctuations are characterized by
           a delayed subtropical gyre response to anomalou's NAO
           surface wind stress forcing and advection of SST
           anomalies originating near the western boundary into the
           interior ocean. Separate ocean-alone experiments suggest
           that the SST variability can not be attributed solely to
           passive response of the ocean to atmospheric thermal
           It is also found that a quasi-oscillatory fluctuation of
           the thermohaline circulation (THC) in the North Atlantic
           ocean with an approximate time scale of 30 years is
           present in the coupled but not in uncoupled simulations.
           The latter were forced with either Newtonian relaxation
           boundary conditions (based on a monthly climatology of
           the atmospheric state variables such as the surface air
           temperature and surface specific humidity) or with
           imposed monthly varying heat and fresh water flux
           These results suggest that the variability of the THC in
           this model is neither an ocean internal phenomenon nor a
           passive response of the ocean to atmospheric forcing.
           Rather, it is a coupled process involving both the ocean
           and the atmosphere. The THC oscillation appears to be
           driven by surface heat flux forcings while the effects of
           surface fresh water fluxes are secondary. Two delay
           effects are crucial for maintaining this oscillation. One
           is the delay of the North Atlantic overturning strength
           relative to the deep water formation rate in the deep
           water production region; the other is the delay of the
           associated SST and surface heat flux anomalies in the
           sinking region relative to the overturning amplitude
           The sea surface temperature signal associated with the
           THC oscillation bears some resemblance to an SST
           interdecadal pattern extracted from observational data
           (Kushnir, 1993). Accompanying anomalies of northern
           hemispheric surface air temperature have positive values
           over the Pacific and Atlantic Oceans and negative values
           over the Eurasian and North American continents. In
           addition to anomalies in sea surface temperature and
           surface air temperature, there are also variations in
           atmospheric flow pattern over the North Atlantic, namely,
           an anomalous northerly flow over the Labrador Sea when
           the THC circulation is strong.
    Title: Modeling the direct and indirect climatic effects of
           tropospheric sulfate aerosols
   Pub No: 9967719
   Author: Cox, Stephen J.
   Degree: PhD
     Date: 2000
    Pages: 103
  Adviser: Wang, Wei-Chyung
     ISBN: 0-599-72439-0
   Source: DAI-B 61/04, p. 1996, Oct 2000
           SCIENCES (0768)
 Abstract: Modeling studies of the climatic effects of tropospheric
           sulfate aerosols are presented. Both the direct
           scattering by the aerosols and the indirect effect of
           enhanced cloud albedo from increased aerosol numbers are
           addressed, in separate studies. The direct effect study
           uses aerosol mass concentrations from the MOGUNTIA
           chemical transport model. A parameterization is developed
           to model the radiative forcing due to direct shortwave
           scattering by the aerosols in the NCAR Community Climate
           Model, an atmospheric general circulation model. Aerosol
           layer optical properties are folded into the direct and
           diffuse surface albedo. The aerosol forcing is similar in
           magnitude, but opposite in sign, to the longwave forcing
           by anthropogenic greenhouse gases such as CO<sub>
           2</sub>, CH<sub>4</sub>, N<sub>2</sub>O,
           CF<sub>2</sub>CL<sub>2</sub>, and CFCL<sub>3</sub>. CCM1
           is run for thirty-five years with equal and opposite
           global annual mean aerosol and greenhouse forcings, and
           the results compared to a control run with no forcing. It
           is determined that the global mean temperate responds to
           the forcings equally, with a global sensitivity of 1.25
           K/(W m<super> &minus;2</super>), but the regional
           temperature response shows marked variation, which could
           not be predicted simply from the forcing pattern. The
           aerosol forcing is concentrated in the industrial
           continental areas of the Northern Hemisphere
           midlatitudes, yet a strong cooling response is noted in
           regions thousands of kilometers away (for instance,
           western Canada) from centers of aerosol concentration.
           The indirect effect is studied with more recent sulfate
           estimates, from the Oslo Chemical Transport Model. Field
           studies are used to relate sulfate mass concentration to
           cloud droplet number concentration, and subsequently to
           cloud droplet effective radius. The indirect
           parameterization is incorporated into NCAR CCM3, along
           with a new shortwave parameterization which allows the
           full vertical distribution of the aerosols to be
           accounted for. The indirect radiative forcing is found to
           be a cooling of 0.47 W m<super>&minus;2</super>, smaller
           than the direct forcing of 1.15 W
           m<super>&minus;2</super>, though still significant. The
           direct forcing is strongest over the Northern Hemisphere
           land, whereas the indirect forcing occurs more strongly
           over the Northern Hemisphere ocean, likely due to greater
           susceptibility of the more pristine marine air to
           condensation nuclei enhancement.
    Title: Longwave radiative transfer in the atmosphere:  Model
           development and applications
   Pub No: 3092290
   Author: Delamere, Jennifer Simmons
   Degree: PhD
     Date: 2003
    Pages: 153
  Adviser: Stamnes, Knut H.
     ISBN: 0-496-39997-9
   Source: DAI-B 64/05, p. 2231, Nov 2003
 Abstract: A FLexible Radiative Transfer Tool (FLRTT) has been
           developed to facilitate the construction of longwave,
           correlated <italic>k</italic>-distribution, radiative
           transfer models. The correlated <italic>k</italic>-
           distribution method is a technique which accelerates
           calculations of radiances, fluxes, and cooling rates in
           inhomogeneous atmospheres; therefore, correlated <italic>
           k</italic>-distribution models are appropriate for
           simulations of satellite radiances and inclusion into
           general circulation models. FLRTT was used to build two
           new rapid radiative transfer models, RRTM_HIRS and
           RRTM_v3.0, which maintain accuracy comparable to the
           line-by-line radiative transfer model LBLRTM.
           <italic>Iacono et al</italic>. [2003] evaluated upper
           tropospheric water vapor (UTWV) simulated by the National
           Center for Atmospheric Research Community Climate Model,
           CCM3, by comparing modeled, clear-sky brightness-
           temperatures to those observed from space by the High-
           resolution Radiation Sounder (HIRS). CCM3 was modified to
           utilize the rapid radiative transfer model RRTM and the
           separate satellite-radiance module, RRTM_HIRS, which
           calculates brightness temperatures in two HIRS channels.
           By incorporating these accurate radiative transfer models
           into CCM3, the longwave radiative transfer calculations
           have been removed as a significant source of error in the
           simulations. An important result of this study is that
           CCM3 exhibits moist and dry discrepancies in UTWV of 50%
           in particular climatic regions, which may be attributed
           to errors in the CCM3 dynamical schemes.
           RRTM_v3.0, an update of RRTM, is a rapid longwave
           radiative transfer appropriate for use in general
           circulation models. Fluxes calculated by RRTM_v3.0 agree
           with those computed by the LBLRTM to within 1.0
           W/m<super>2</super> at all levels, and the computed
           cooling rates agree to within 0.1 K/day and 0.3 K/day in
           the troposphere and stratosphere, respectively.
           This thesis also assessed and improved the modeling of
           clear-sky, longwave radiative fluxes at the Atmospheric
           Radiation Measurement Program North Slope of Alaska site
           by simultaneously addressing the specification of the
           atmosphere, radiometric measurements, and radiative
           transfer modeling. Consistent with findings from other
           field sites, the specification of the atmospheric water
           vapor is found to be a large source of uncertainty in
           modeled radiances and fluxes. Improvements in the
           specification of carbon dioxide optical depths within
           LBLRTM resulted, in part, from this analysis.
    Title: Speciated local aerosol characteristics and radiative
           forcing at a rural midwestern site
   Pub No: 9971067
   Author: Dillner, Ann Marie
   Degree: PhD
     Date: 2000
    Pages: 162
  Adviser: Larson, Susan M.
     ISBN: 0-599-76207-1
   Source: DAI-B 61/05, p. 2697, Nov 2000
           SCIENCE (0608)
 Abstract: In this research, physical and chemical properties of
           ambient aerosols were measured at a rural perturbed mid-
           latitude site (Bondville, IL) and used to calculate the
           aerosol optical properties and the resulting direct
           radiative forcing. Size-segregated aerosol samples were
           collected during the summer of 1997 using three parallel
           MOUDIs operating at ambient relative humidity. Two sample
           sets were used to obtain sulfate, organic carbon (OC),
           elemental carbon (EC), carbonate and total aerosol mass.
           The third sample set was used to obtain the size-specific
           and wavelength-dependent extinction efficiency of EC.
           The measured submicrometer mass concentration was 11.4
           &plusmn; 4.0 &mu;g m<super>&minus;3</super>. Ammonium
           sulfate comprised nearly half of the submicrometer
           aerosol and OC plus EC comprised 25%.
           Water content for ammonium sulfate and OC was estimated
           using both K&ouml;hler theory and parameterized water
           uptake curves from the literature. Water content for
           internally mixed aerosols was determined using a ZSR
           method. Aerosol optical properties (extinction
           efficiency, asymmetry parameter, single scatter albedo)
           were calculated from measured size distributions and
           wavelength dependent refractive indexes for each species
           and for internal and external mixtures using Mie theory.
           A technique, utilizing transmission measurements through
           extracts of size segregated ambient aerosol samples, was
           developed to obtain the extinction efficiency of EC.
           Measured EC extinction efficiencies ranged from 7.3 to
           1.7 m<super>2</super> g<super>&minus;1</super> at 550 nm,
           depending on particle diameter. Normalized direct aerosol
           radiative forcing (W g<super>&minus;1 </super>) was
           calculated using the Column Radiation Module (CRM) of the
           National Center for Atmospheric Research (NCAR) Community
           Climate Model (CCM3). Aerosol optical properties, used in
           the model, were assumed to be uniform throughout the
           lowest one kilometer of the atmosphere. The normalized
           forcing due to ammonium sulfate was &minus;340 &plusmn;
           10 W g<super>&minus;1</super>. OC was 1/3 larger and
           residue was 1/3 smaller. EC within an internally mixed
           aerosol was shown to suppress forcing.
           This research supports the view that species other than
           ammonium sulfate, namely EC, OC and dust-like aerosol,
           have significant radiative effects. Although the mass
           percentage of EC may be low at rural sites, ignoring the
           presence of EC aerosol leads to an over estimation of the
           aerosol forcing.
    Title: Short-range ensemble forecasting of an explosive
           cyclogenesis with a limited-area model
   Pub No: 9626568
   Author: Du, Jun
   Degree: PhD
     Date: 1996
    Pages: 146
  Adviser: Mullen, Steven L.; Sanders, Frederick
   Source: DAI-B 57/04, p. 2623, Oct 1996
 Abstract: Since the atmosphere is a chaotic system, small errors in
           the initial condition of any numerical weather prediction
           (NWP) model amplify as the forecast evolves. To estimate
           and possibly reduce the uncertainty of NWP associated
           with initial-condition uncertainty (ICU), ensemble
           forecasting has been proposed which is a method of,
           differently from the traditional deterministic
           forecasting, running several model forecasts starting
           from slightly different; initial states.
           In this dissertation, the impact of ICU and short-range
           ensemble forecasting (SREF) on quantitative precipitation
           forecasts (QPFs), as well as on sea-level cyclone
           position and central pressure, is examined for a case of
           explosive cyclogenesis that occurred over the contiguous
           United States. A limited-area model (the PSU/NCAR NM4) is
           run at 80-km horizontal resolution and 15 layers to
           produce a 25-member, 36-h forecast ensemble. Lateral
           boundary conditions for the MM4 model are provided by
           ensemble forecasts from a global spectral model (the NCAR
           CCM1). The initial perturbations of the ensemble members
           possess a magnitude and spatial decomposition which
           closely match estimates of global analysis error, but
           they were not dynamically-conditioned. Results for 80-km
           ensemble forecast are compared to forecasts from the then
           operational Nested Grid Model (NGM), a single 40-km MM4
           forecast, and a second 25-member MM4 ensemble based on a
           different cumulus parameterization and slightly different
           initial conditions.
           Acute sensitivity to ICU marks ensemble QPF and the
           forecasts of cyclone position and central pressure.
           Ensemble averaging always reduces the rms error for QPF.
           Nearly 90% of the improvement is obtainable using
           ensemble sizes as small as 8-10. However, ensemble
           averaging can adversely affect the forecasts related to
           precipitation areal coverage because of its smoothing
           nature. Probabilistic forecasts for five mutually
           exclusive, completely exhaustive categories are found to
           be skillful relative to a climatological forecast.
           Ensemble sizes of $/sim$10 can account for 90% of
           improvement in probability density function.
           Our results indicate that SREF techniques can now provide
           useful QPF guidance and increase the accuracy of
           precipitation, cyclone position, and cyclone's central
           pressure forecasts. With current analysis/forecast
           systems, the benefit from simple ensemble averaging is
           comparable to or exceed that obtainable from improvement
           in the analysis/forecast system.
    Title: Regional simulation of North American interannual climate
   Pub No: 9998321
   Author: Dutton, Jan Frederik
   Degree: PhD
     Date: 2000
    Pages: 134
  Adviser: Barron, Eric
     ISBN: 0-493-06588-1
   Source: DAI-B 61/12, p. 6518, Jun 2001
 Abstract: The interannual variability of the RegCM2 regional
           climate model, with a U.S. centered domain, forced by the
           NCAR CCM3 global climate is the main focus of this study.
           A 6 member 10 year CCM3 ensemble, from January 1, 1968 to
           December 31, 1978, forced globally by reconstructed
           observed sea surface temperatures, provides RegCM2
           boundary conditions.
           An understanding of how RegCM2 responds to CCM3
           interannual variability is developed using an anomaly
           pattern correlation (APC) analysis of the CCM3 and RegCM2
           monthly averaged 500 mb heights, surface temperature, and
           precipitation. All three variables show a distinct APC
           annual cycle.
           The effect of regional climate modeling on simulated
           climate reproducibility compared using the modeled
           ensemble run variances. In certain regions RegCM2 surface
           temperature normalized ensemble run variance is
           significantly lower than CCM3, suggesting higher
           reproducibility in these regions. RegCM2 precipitation
           reproducibility is found to be lower than CCM3 in all
           regions, indicating higher resolution simulations have
           lower reproducibility.
           A empirical orthogonal function (EOF) analysis of the
           dominant precipitation and surface temperature modes of
           variability is also presented. The EOF analyses use 60
           years (6 ensemble members x 10 years) of RegCM2, CCM3,
           and a gridded observed data. RegCM2 does not
           significantly alter the primary modes of CCM3 surface
           temperature variability. Furthermore, the observed
           primary modes of surface temperature variability are
           similar to those modeled. Neither model captures the
           primary mode of winter precipitation variability. The
           modeled primary mode of precipitation variability,
           focused on the west U.S. coast, is similar to the
           observed second mode. The RegCM2 increases the amplitude
           of the CCM3 mode, improving the simulation of winter
           precipitation variability relative to the observations.
           The analysis of summer surface temperature and
           precipitation variability reveals that regional scale
           forcing, such as the low-level jet, plays an important
           role in modifying the locations of the primary modes of
           variability. The RegCM2 significantly improves the jet
           simulation, shifting the precipitation variability to
           midwestern and eastern U.S. Despite these changes, the
           results show the CCM3 interannual variability signal is
           present in the RegCM2 signal throughout the year.
    Title: Tropical climate stability, Hadley circulation, and deep
           cumulus convection:  Vital synergism on a wet planet
   Pub No: 9975794
   Author: Fleischfresser, Luciano
   Degree: PhD
     Date: 2000
    Pages: 129
  Adviser: Fiedler, Brian
     ISBN: 0-599-81465-9
   Source: DAI-B 61/06, p. 3095, Dec 2000
           SCIENCES (0768); APPLIED MECHANICS (0346)
 Abstract: This research is about developing a more internally
           consistent formulation of cumulus convection and the
           atmospheric branch of the hydrological cycle in a newly
           developed Hadley circulation model. The ultimate goal is
           to analyze the climate equilibrium of a symmetric
           tropical hydrostatic atmosphere, particularly studying
           cause and effect relations determining the magnitude of
           water vapor-related feedbacks in climate sensitivity
           Important features of the proposed formulation for
           precipitating deep turbulent clouds include the ability
           to calculate precipitation efficiencies, a new
           postulation relating cumulus buoyancy to solar radiation,
           and the implicit account of latent heat release that
           manifest itself by scaling cumulus drafts to observed
           magnitudes. The Hadley model is based on primitive
           equations, and it incorporates the NCAR column radiation
           scheme as well as the atmospheric branch of the
           hydrological cycle. To calculate climatic feedbacks, the
           inverse simulation approach is used. An interrelationship
           technique is applied to diagnose feedback factors
           associated with changes of water vapor amount and
           distribution, of lapse-rate, and of deep cumulus cloud
           cover. The aim is to contrast the novel model for deep
           clouds with a mass flux deep cumulus parameterization
           when a thermally direct circulation (Hadley cell) is
           When only a lapse-rate adjustment is used to crudely
           represent tropical convection, the calculated climate
           sensitivity lies in the typical range of equatorial
           sensitivities given by global circulation models (GGMs).
           This result suggests that the Hadley model is capturing
           the essential physics of these models as far as these
           sensitivities are concerned. In the comparison analyses,
           the climate equilibrium is stable and effected by net
           positive feedback when the new cumulus model is used.
           Moreover, the calculated tropical climate sensitivities
           are consistently lower than the aforementioned typical
           range, bringing them closer to sensitivities suggested by
           observed data. Interestingly, the tropical climate
           equilibrium is unstable with the mass flux scheme. It is
           shown quantitatively that relative humidity changes in
           the model upper-troposphere determine the sign of the
           water vapor feedback. Recommendations to narrow humidity
           uncertainties in climate change simulations are
    Title: The development of a three-dimensional general
           circulation model with coupled chemistry
   Pub No: 9805287
   Author: Gross, Gerhard Wayne
   Degree: PhD
     Date: 1997
    Pages: 382
  Adviser: Khalil, M. Aslam K.
     ISBN: 0-591-55236-1
   Source: DAI-B 58/08, p. 4282, Feb 1998
 Abstract: A coupled climate-chemistry model for the troposphere and
           lower stratosphere has been developed that includes on-
           line local actinic flux calculation. The coupling of
           chemistry with a climate model and the internal
           calculation of actinic flux allows the model to simulate
           a variety of interactions and feedbacks between
           atmospheric processes that are ordinarily excluded from
           general circulation models.
           The climate model is an altered version of the CCM3
           (Community Climate Model Version 3) from NCAR (National
           Center for Atmospheric Research). The spectral resolution
           of the radiation code of the model has been increased 8-
           fold in the region 200-700 nm, with highest resolution
           between 245-350 nm. Actinic flux is calculated by the
           radiation code for each spectral interval and each model
           grid cell at a user defined frequency set to 1 hour in
           this work. The coupled tropospheric chemistry component
           uses climatological data from the climate model including
           temperature, pressure, water vapor concentration, actinic
           flux, and atmospheric transport.
           The OH distributions generated by a 3 year simulation of
           the coupled model in the partially coupled mode show a
           zonal maximum over the tropics as expected. Seasonal
           variations displace the zonal maximum towards the summer
           pole by about 5$/sp/circ$ in the southern hemisphere and
           about 15$/sp/circ$ in the northern hemisphere. The
           distributions show somewhat more structure compared to
           results from other model studies. During the solstice
           months the OH concentration is greatest at the surface
           and during the equinox months the maximum is slightly
           elevated to approximately 1.5 km. The concentration of
           the equatorial maximum decreases with altitude and then
           increases above about 400 mb (7 km).
           The model calculated, global, tropospheric, average is
           1.1 $molecules/cm/sp3$ and is slightly higher than
           results from other studies. The global average at the
           surface is 1.2 $molecules/cm/sp3$ and agrees with
           measurements and model studies. The globally averaged
           amount of $CH/sb4$ destroyed by oxidation with OH is
           calculated to be 461 $Tg/yr$ and also agrees with
           independent estimates from other researchers. The
           distribution of formaldehyde is consistent with
           measurements made by other groups.
    Title: I. Variability of the outgoing thermal IR spectra and its
           application in GCM validation. II. The detection of
           cloud/aerosol in the outgoing thermal IR spectra
   Pub No: 3151366
   Author: Huang, Xianglei
   Degree: PhD
     Date: 2004
    Pages: 182
  Adviser: Yung, Yuk L.
     ISBN: 0-496-11390-9
   Source: DAI-B 65/10, p. 5193, Apr 2005
           AND ASTROPHYSICS (0606)
 Abstract: The theme of this thesis is studying the outgoing
           infrared spectra of Earth and Mars. In Chapter 1, the
           importance and the feasibility of using the outgoing
           thermal IR spectra in testing model's variability are
           discussed. Chapter 2 investigates the temporal
           variability over the tropical and midlatitude Pacific
           seen from Infrared Interferometer Spectrometer (IRIS)
           observations and corresponding synthetic spectra based on
           simulations from two general circulation models (GCMs),
           UCLA GCM and NCAR CAM2. The discrepancies between modeled
           and observed variability are substantial. Further
           examination shows that these discrepancies are most
           likely due to deficiencies in simulating the seasonal and
           intraseasonal variations of the Walker Circulation in the
           tropical Pacific and the seasonal variations of low
           cloud, boundary-layer temperature, and stratospheric
           temperature in the midlatitude Pacific. Chapter 3
           presents a survey of the spatial variability seen from
           Atmospheric Infrared Sounder (AIRS) spectra and
           corresponding synthetic spectra based on NCAR CAM2
           simulation. To a large extent, the simulated variability
           agrees well with the observed counterpart. The major
           discrepancies can be attributed to the incorrect location
           of ITCZ in the western Pacific, the underrepresented dust
           aerosol over the Arabian Sea and off the Atlantic Coast
           of North Africa, and the overestimated spatial variation
           of stratospheric temperature in the model. Chapter 4
           presents a comparative study of the variability seen in
           the Martian outgoing infrared spectra collected by MGS-
           TES (Thermal Emission Spectrometer). The negative
           correlation between dust and water ice spectral features
           seen in this analysis suggests that, to some extent, dust
           and water ice cloud are mutually exclusive of each other
           in the Martian atmosphere. Chapter 5 presents a
           sensitivity study of identifying optically thin cirrus
           from high-resolution infrared spectra based on the line
           shapes in the residual spectra. This cirrus-detection
           approach is original in the sense of making use of
           information content contained in such measurements.
           Chapter 6 presents a tri-spectral algorithm to detect
           water ice cloud, dust, and surface anisothermality from
           low-resolution Martian outgoing thermal IR spectra. This
           algorithm can be used to screen large amounts of data to
           get a quick overview.
    Title: Adaptive grids in weather and climate modeling
   Pub No: 3121951
   Author: Jablonowski, Christiane
   Degree: PhD
     Date: 2004
    Pages: 266
  Adviser: Penner, Joyce E.
     ISBN: 0-496-69302-7
   Source: DAI-B 65/02, p. 792, Aug 2004
 Abstract: Adaptive Mesh Refinement (AMR) provides an attractive
           framework for atmospheric flows since it allows improved
           horizontal resolution in a limited region without
           requiring a fine grid resolution throughout the entire
           model domain. In this thesis, the adaptive grid technique
           has been applied to a revised version of NCAR/NASA's next
           generation dynamical core for climate and weather
           research. This hydrostatic so-called Lin-Rood dynamics
           package with a conservative finite volume discretization
           in flux form provides highly efficient algorithms for
           high performance computing.
           The adaptive model design utilizes a spherical adaptive-
           grid library which is based on a cache-efficient block-
           structured data layout. This AMR communication library
           for parallel processors has been newly developed in the
           Computer Science Department at the University of
           Michigan. All blocks are self-similar and split into four
           in the event of refinement requests. The resolution of
           neighboring blocks can only differ by a factor of two
           which leads to cascading refinement regions.
           The adaptive dynamical core is run in two configurations:
           the full 3D hydrostatic dynamical core on the sphere and
           the corresponding 2D shallow water model that has been
           extracted out of the 3D version. This shallow water setup
           serves as an ideal testbed for the horizontal
           discretization and the 2D adaptive-mesh strategy. It
           further allows the efficient testing of interpolation
           routines at fine-coarse grid interfaces.
           The static and dynamic adaptations are tested using the
           standard shallow water test suite and a newly-developed
           idealized 3D baroclinic wave test case. Static
           adaptations are used to vary the resolution in pre-
           defined regions of interest. This includes static
           refinements near mountain ranges or static coarsenings in
           the longitudinal direction for the implementation of a
           so-called reduced grid in polar regions. Dynamic
           adaptations are based on flow characteristics and guided
           by refinement criteria that detect user-defined features
           of interest during a simulation. In particular, flow-
           based refinement criteria, such as vorticity or gradient
           indicators, are suggested. Refinements and coarsenings
           occur according to pre-defined threshold values.
           This research project is characterized by an
           interdisciplinary approach involving atmospheric science,
           computer science and applied mathematics.
    Title: A physically-based snow model coupled to a general
           circulation model for hydro-climatological studies
   Pub No: 3050315
   Author: Jin, Jiming
   Degree: PhD
     Date: 2002
    Pages: 162
  Adviser: Sorooshian, Soroosh
     ISBN: 0-493-64911-5
   Source: DAI-B 63/04, p. 1761, Oct 2002
 Abstract: A Snow-Atmosphere-Soil Transfer (SAST) model has been
           developed to extend the point snowmelt model to vegetated
           areas using the parameterization concepts of the
           Biosphere-Atmosphere Transfer Scheme (Dickinson et al.
           1993). The model applications for short-grass and forest
           fields show that the simulated surface temperature,
           albedo, and snow depth have close agreement with
           observations. In addition, because of biases in simulated
           runoff in the high-latitudes, a Shuffled Complex
           Evolution (Sorooshian et al. 1993) scheme for automatic
           calibration has been connected with the SAST model to
           determine the realistic distribution of runoff components
           from different soil layers and search the optimized
           parameter set. The calibrated runoff closely matches
           Because the Community Climate Model version 3 (CCM3)
           coupled with the SAST model overestimates snow depth and
           precipitation and underestimates surface temperature over
           the Rocky Mountains, remotely sensed snow depth data have
           been assimilated in the model to alleviate model
           discrepancies based on energy and mass balances. The
           improved surface temperature simulations result from the
           decreased snowmelt and albedo in winter and spring and
           from the weakened evaporation in summer due to drier
           soil. Meanwhile, modeled summer precipitation over the
           Rocky Mountains has a minor improvement.
           The relationship between the variations of tropical
           Pacific SST and snowpack anomalies in the western United
           States (U.S.) has been studied by comparing observations
           and CCM3 output. The results indicate that in the
           northwestern U.S., the warm tropical Pacific phase of the
           El Ni&ntilde;o-Southern Oscillation (ENSO) is associated
           with diminished snowpack while its cool phase is related
           to enhanced snowpack. This relationship is largely
           determined by winter precipitation variability for the
           observations; however, it relies heavily on the
           variations of temperature due to the biases in
           atmospheric patterns for the model output. In the
           southwestern U.S., positive snowpack anomalies for both
           observations and simulations result from the strong warm
           phase of the ENSO and negative ones are connected with
           exaggerated local precipitation in fall.
    Title: Interpolation of surface radiative temperature measured
           from polar orbiting satellites to a diurnal cycle
   Pub No: 9923351
   Author: Jin, Menglin
   Degree: PhD
     Date: 1999
    Pages: 152
  Adviser: Dickinson, Robert E.
     ISBN: 0-599-22782-6
   Source: DAI-B 60/03, p. 1128, Sep 1999
           (0799); ENVIRONMENTAL SCIENCES (0768)
 Abstract: The land surface skin temperature diurnal cycle (LSTD) is
           very important for the understanding of surface climate
           and for evaluating climate models. This variable,
           however, cannot be obtained globally from polar-orbiting
           satellites because the satellites usually pass a given
           area twice per day and because their infrared channels
           cannot observe the surface when the sky is cloudy.
           In order to more optimally use the satellite data, this
           research is designed, for the first time, to solve the
           above two problems by advance use of remote sensing
           techniques and climate modeling. Specifically, this work
           is divided into two parts. Part one deals with obtaining
           the skin temperature diurnal cycle for cloud-free cases.
           We have developed a &ldquo;cloud-free algorithm&rdquo; to
           combine model results with satellite and surface-based
           observations, thus interpolating satellite twice-daily
           observations to the diurnal cycle. Part two studies the
           cloudy cases. The &ldquo;cloudy-pixel treatment&rdquo;
           presented here is a hybrid technique of
           &ldquo;neighboring-pixel&rdquo; and &ldquo;surface air
           temperature&rdquo; approaches. The whole algorithm has
           been tested against field experiments and climate model
           CCM3/BATS in global and single column mode simulations.
           It shows that this proposed algorithm can obtain skin
           temperature diurnal cycles with an accuracy of 1&ndash;2
           K at the monthly pixel level.
    Title: The onset of convection in the Madden-Julian oscillation
   Pub No: 9948702
   Author: Kemball-Cook, Susan Rives
   Degree: PhD
     Date: 1999
    Pages: 109
  Adviser: Weare, Bryan
     ISBN: 0-599-51118-4
   Source: DAI-B 60/10, p. 5117, Apr 2000
 Abstract: An observational study of the onset of convection in the
           Madden-Julian Oscillation (MJO) was performed. Composites
           of radiosonde data from the Comprehensive Aerological
           Reference Data Set were constructed for an ensemble of
           tropical stations in the Indian Ocean, Maritime
           Continent, and western Pacific Ocean. Outgoing longwave
           radiation from NOAA polar orbiting satellites was used as
           a proxy for deep convection associated with the MJO, and
           NCEP-NCAR Reanalysis data was used to diagnose low-level
           The composites suggest that for the off-equatorial
           stations used in this study, the MJO period may be set by
           the buildup and discharge of the low-level moist static
           energy (h). This result supports the discharge-recharge
           hypothesis of Blade and Hartmann (1993). MJO events are
           most likely to occur when the off-equatorial atmosphere
           has been destabilized through a combination of low level
           h buildup and concurrent drying of the middle troposphere
           by subsidence in the wake of the previous cycle of MJO
           convection. The low-level h buildup is controlled by a
           corresponding increase in low-level moisture. Following
           an episode of MJO convection, a period of approximately
           35 days is required to increase the low-level moisture to
           the point where the atmosphere again becomes unstable to
           deep convection. The convective event lasts about 20 days
           and stabilizes the atmosphere by drying and cooling the
           boundary layer and moistening and heating the middle and
           upper troposphere. By drying and cooling the boundary
           layer, convection discharges the low level h.
           The increase in low-level moisture is not caused by the
           1000 mb convergence, as suggested by frictional wave-CISK
           theories. For the stations examined here, the convergence
           lags the moist static energy buildup, and is instead in
           phase with the convection. A possible mechanism for the
           relatively slow moisture buildup lasting approximately 35
           days is competition between surface fluxes of moisture
           and the drying effect of entrainment of low h air into
           the boundary layer. Evidence is found for the
           preconditioning of the atmosphere to deep convection
           through upward transport of moisture by shallow
    Title: Tropical dynamics near the stratopause: The two-day wave
           and its relatives
   Pub No: 9836209
   Author: Limpasuvan, Varavut
   Degree: PhD
     Date: 1998
    Pages: 209
  Adviser: Leovy, Conway B.
     ISBN: 0-591-89715-6
   Source: DAI-B 59/06, p. 2808, Dec 1998
           SCIENCES (0768)
 Abstract: The two-day wave is observed in the Upper Atmosphere
           Research Satellite (UARS) Microwave Limb Sounder (MLS)
           temperature and water vapor data. During a 3-year period
           (Dec. 1991-Sep. 1994), the wave signature is prominent
           semiannually after each solstice and is comprised of a
           zonal wavenumber 3 component with $[/sim]2.0$-day period
           and a wavenumber 4 component with $[/sim]1.8$-day period.
           Intra-annually, the wavenumber 3 amplitude during the
           austral summer is nearly twice as strong as during the
           boreal summer. The wavenumber 4 component however can be
           equally strong in both summers. The wave-number 3 signal
           tends to be dominant during the austral summer while the
           wave 4 component is dominant during the boreal summer.
           The derived zonal wind structure suggests that the
           Charney-Stern inflection instability mechanism is
           responsible for generating the two-day wave whose
           amplitude resides mainly on the equatorward flank of the
           summer easterly jet.
           In some seasons, momentum redistribution by inertial
           instability appears to destabilize the easterly jet from
           which the barotropically unstable two-day wave grows. In
           these instances, an independent study using the UARS
           Cryogenic Limb Array Etalon Spectrometer (CLAES)
           temperature (Hayashi et al., 1998) coincidently
           identifies 'pancake' structures associated with inertial
           instability. A possible connection between inertial
           instability and the two-day wave has been discussed by
           Hitchman (1985) and Orsolini et al. (1997).
           The National Center for Atmospheric Research (NCAR)
           Community Climate Model version 2 (CCM2) mechanistic
           model is used to simulate this connection. Model
           experiments show that, for a prescribed initial wind
           condition with strong horizontal wind shear in the
           tropics, inertial instability can trigger the two-day
           wave. An increase in easterly wind curvature fostered by
           inertial instability circulation destabilizes the jet in
           low summer latitudes and allows first the growth of the
           wavenumber 4 then the wavenumber 3 component of the two-
           day wave. Near the stratopause, the two-day wave energy
           propagation is directed equatorward away from the wave's
           critical line source and westerly momentum is fluxed into
           the easterly jet core by the wave. While much of the
           wavenumber 4 activity is confined near the stratopause,
           the wavenumber 3 energy can propagate upward well into
           the mesosphere where strong Rayleigh damping is imposed.
           The simulated waves' spatial and spectral characteristics
           are fairly realistic.
    Title: On the relationship between tropical convection and clear
           sky upper troposphere moisture
   Pub No: 3088067
   Author: Lin, Wuyin
   Degree: PhD
     Date: 2002
    Pages: 103
  Adviser: Zhang, Minghua
     ISBN: 0-496-35829-7
   Source: DAI-B 64/04, p. 1770, Oct 2003
 Abstract: Upper troposphere water vapor is a crucial factor in the
           climate system. The moisture abundance in the clear sky
           upper troposphere is strongly influenced by tropical deep
           convection. TOVS upper troposphere water vapor
           measurements and ISCCP deep convective clouds are used to
           investigate the relationship between upper troposphere
           humidity (UTH) and frequency and area coverage of
           tropical deep convection. The study shows that a more
           diffused distribution of convection in the tropics is
           associated with a higher UTH in the free troposphere,
           while a more concentrated convection distribution is
           associated with a drier free upper troposphere. It is
           also shown that increased convective area coverage is
           accompanied with an increased area coverage of less
           frequent convection and a decreased area coverage of
           highly frequent convection, and vice versa. The
           identification of the geographical distribution of
           convection perturbation associated with a potential
           climate change is therefore important to the upper
           troposphere water vapor feedback.
           The mechanism behind the observed relationship is
           interpreted using trajectory analysis. The results show
           that a more diffused distribution of convection gives
           rise to a shorter average horizontal distance and a less
           vertical displacement between convection and clear
           regions, and hence a moister clear sky condition.
           Similarly, when deep convection is more concentrated, the
           average traveling distance is longer, and the clear sky
           upper troposphere is drier.
           This process is verified using a simple heating driven
           circulation model with prescribed distributions of
           convective latent heating. The NCAR CCM is then used to
           examine the relationship between UTH and convection. The
           model is found to capture the relationship between water
           vapor concentration and total convective area reasonably,
           but it fails to reproduce the respective contributions
           from highly and less frequent convections. The deficiency
           lies in the fact that convection frequency is broadly
           overestimated in the model. The area coverages of
           different convective regimes are not simulated correctly.
           The increase of total spatial coverage of deep convection
           in the model is mainly reflected in the increased area
           coverage of highly frequent convection, leading to an
           opposite relationship between UTH and highly frequent
           convection compared with the observations.
    Title: Improvement in runoff parameterization for global climate
   Pub No: NQ58635
   Author: Liu, Jinliang
   Degree: PhD
     Date: 2001
    Pages: 168
  Adviser: Cho, Han-Ru
     ISBN: 0-612-58635-9
   Source: DAI-B 62/04, p. 1906, Oct 2001
 Abstract: Diagnosed runoff was estimated from NCEP/NCAR reanalysis
           data for an 8-year period from 1987 to 1994. Bonan's land
           surface model (LSM) was run for the same period coupled
           to NCAR's CCM3. Comparisons between the diagnosed and
           simulated runoff indicate that, the runoff
           parameterization in the original LSM cannot produce a
           reasonable diagnosed horizontal distribution of runoff,
           which is important to the model climate. One of the
           possible reasons is the exclusion of topographic effects
           in the original runoff parameterization. Therefore, based
           on previous research results on river re-distribution
           models, a modification on the original runoff
           parameterization was proposed and implemented in the
           original LSM. This modification has two aspects: firstly,
           the topographic slopes cause outflows from higher
           topography and inflows into the lower topography points;
           secondly, topographic slopes also cause decrease of
           infiltration at higher topography and increases of
           infiltration at lower topography. Then changes in
           infiltration result in changes in soil-moisture, surface
           fluxes and then in surface temperature. This mechanism is
           very clearly demonstrated in the point budgets analysis
           at the Andes Mountains vicinities. Analysis from a re
           gional scale perspective in the Canadian GEWEX basin, the
           Mackenzie River basin, shows that the modified runoff
           parameterization can generate an expected horizontal
           distribution of total runoff which is much closer to the
           horizontal distribution of the observed and diagnosed
           runoff, and which is much more consistent with topography
           and thus very easy to explain physically. This represents
           a significant improvement over the original LSM. More
           importantly, very detailed analysis from a global
           perspective shows many very encouraging improvements
           introduced by the modified model over the original model
           in simulating basic atmospheric climate properties such
           as thermodynamic features, precipitable water, net water
           exchange, and precipitation. The modified model somehow
           corrected some deficiencies evident in the AGCM. All of
           these improvements in the atmospheric climate simulation
           illustrate that the inclusion of topographic effects in
           the LSM can force the AGCM to produce a more realistic
           model climate. Analysis also shows that the modified
           model may improve the atmospheric <italic>CO
           </italic><sub>2</sub> simulation which is very important
           to global environmental studies.
    Title: Sea ice climatology, variations and teleconnections:
           Observational and modeling studies
   Pub No: 3080822
   Author: Liu, Jiping
   Degree: PhD
     Date: 2003
    Pages: 140
  Adviser: Martinson, Douglas G.; Rind, David H.
     ISBN: 0-496-28696-6
   Source: DAI-B 64/02, p. 768, Aug 2003
           OCEANOGRAPHY (0415); GEOPHYSICS (0373)
 Abstract: Hypotheses, models and observations suggest that sea ice
           plays an important role in the local, regional and global
           climate through a variety of processes across a full
           range of scales. However, our documentation and
           understanding of the nature of the polar-extrapolar
           climate teleconnections and their underlying causal and
           mechanistic links are still rudimentary, and the largest
           disagreements among model simulations of present and
           future climate are in the polar regions.
           In an effort to address these issues, we evaluated the
           simulated Antarctic sea ice variability and its climate
           teleconnections in three coupled global climate models
           (GISS, NCAR and GFDL) as compared to the observations.
           All the models capture the El Ni&ntilde;o-Southern
           Oscillation (ENSO)-like phenomenon to some degree,
           although almost all the models miss some observed
           linkages. The GISS and NCAR models also capture the
           observed Antarctic Dipole and meridional banding
           structure through the Pacific. The Antarctic sea ice
           regions showing the strongest sensitivity to global
           teleconnections differ among the models and between the
           models and observations. We then proposed that the
           changes of the regional mean meridional atmospheric
           circulation (the regional Ferrel Cell) are one such
           mechanism leading to the covariability of the ENSO and
           Antarctic Dipole by modulating the mean meridional heat
           flux using the observational data.
           To more accurately represent sea ice simulations and
           associated feedbacks with the atmosphere and the ocean,
           the GISS coupled model was used to investigate the
           sensitivity of sea ice to the following physical
           parameterizations: (a)&nbsp;two sea ice dynamics
           (cavitating fluid and viscous-plastic), (b)&nbsp;the
           specification of oceanic isopyncal mixing coefficients in
           the Gent and McWillams isopyncal mixing, (c)&nbsp;the
           Wajsowicz viscosity diffusion, (d)&nbsp;surface albedo,
           (e)&nbsp;the penetration of solar radiation in sea ice,
           (f)&nbsp;effects of including a sea ice salinity budget,
           and (g)&nbsp;the ice-ocean boundary interactions. The
           atmospheric responses associated with sea ice changes
           were discussed. Based on these experiments, a series of
           composite experiments with the aforementioned
           parameterizations were also investigated. While
           improvements are seen overall, there are some unrealistic
           aspects that will require further improvements to the sea
           ice and ocean components. Finally, the GISS coupled
           model's ability to represent observed Antarctic sea ice
           variability and its global teleconnections was re-
           evaluated in a composite run with the improved sea ice
           and ocean processes.
    Title: Parameter estimation for locally coupled land surface-
           atmosphere models
   Pub No: 3108924
   Author: Liu, Yuqiong
   Degree: PhD
     Date: 2003
    Pages: 239
  Adviser: Sorooshian, Soroosh
     ISBN: 0-496-56469-8
   Source: DAI-B 64/10, p. 4833, Apr 2004
  Subject: HYDROLOGY (0388)
 Abstract: As land-surface modeling moves from the off-line mode to
           the coupled mode, it is also highly desirable to extend
           the off-line calibration of land-surface models to
           coupled applications. Using the NCAR SCCM as an example,
           this study proposed and implemented some effective
           schemes for the application of automatic parameter
           estimation procedures in a locally coupled environment,
           where other relevant issues such as parameterization
           tests, sensitivity analyses, and off-line calibrations
           were also involved.
           A parameterization deficiency having serious negative
           impacts on the performance of the NCAR SCCM was
           identified and rectified in this work, which led to
           significantly improved model performances and formed the
           basis for the subsequent sensitivity analysis and
           calibration experiments. To facilitate the calibration
           studies, an independent sensitivity analysis was
           conducted to identify some sensitive model parameters,
           followed by a multi-objective sensitivity analysis using
           the MOGSA algorithm to obtain better understanding of the
           model. Some off-line calibrations using the NCAR LSM were
           also conducted for comparison purposes.
           In the locally coupled environment, both land-surface and
           atmospheric variables/parameters were involved in the
           calibration processes of 14 different pre-designed
           calibration cases. In brief, the results show that
           atmospheric parameters are of critical importance for the
           calibration of a coupled land surface-atmosphere model,
           and atmospheric forcing variables generally contain more
           useful information for calibration than land-surface
           fluxes/variables. In the coupled environment, step-wise
           calibration schemes, with land-surface and atmospheric
           parameters optimized successively in the off-line and
           coupled modes, respectively, appear to be superior to the
           single-step calibration schemes which optimize land-
           surface and atmospheric parameters simultaneously in the
           coupled environment, in that the former can provide
           better converged optimal solutions with less
           uncertainties. In addition, the results also show that
           better optimization effects can be achieved in the
           partially decoupled environment by replacing the model-
           generated precipitation and net radiation with the
           corresponding observations to drive the land-surface part
           of the model, indicating the dominant importance of
           precipitation and radiation in a coupled land surface-
           atmosphere model.
    Title: Applications of tree-structured regression for regional
           precipitation prediction
   Pub No: 9987703
   Author: Li, Xiangshang
   Degree: PhD
     Date: 2000
    Pages: 212
  Adviser: Sailor, D. J.
     ISBN: 0-599-94468-4
   Source: DAI-B 61/09, p. 4779, Mar 2001
 Abstract: This thesis presents a Tree-Structured Regression (TSR)
           method to relate daily precipitation with a variety of
           free atmosphere variables. Historical data were used to
           identify distinct weather patterns associated with
           differing types of precipitation events. Models were
           developed using 67% of the data for training and the
           remaining data for model validation. Seasonal models were
           built for each of four U.S. sites; New Orleans Louisiana,
           San Antonio and Amarillo of Texas as well as San
           Francisco California. The average correlation by site
           between observed and simulated daily precipitation data
           series range from 0.69 to 0.79 for the training set, and
           0.64 to 0.79 for the validation set. Relative humidity
           related variables were found to be the dominant variables
           in these TSR models. Output from an NCAR Climate System
           Model (CSM) transient simulation of climate change were
           then used to drive the TSR models for predicting
           precipitation characteristics under climate change. A
           preliminary screening of the GCM output variables for
           current climate, however, revealed significant problems
           for the New Orleans, San Antonio and Amarillo sites.
           Specifically, the CSM missed the annual trends in
           humidity for the grid cells containing these sites. CSM
           output for the San Francisco site was found to be much
           more reliable. Therefore, we present future precipitation
           estimates only for the San Francisco site. While both GCM
           and TSR predict very small change in overall annual
           precipitation, they differ significantly from season to
    Title: Effects of cloud-radiative forcing on atmospheric
           response to tropical SST anomaly: Observation and
   Pub No: 9712812
   Author: Lu, Xiaodan
   Degree: PhD
     Date: 1996
    Pages: 134
     ISBN: 0-591-20372-3
   Source: DAI-B 57/11, p. 6827, May 1997
           SCIENCE (0608); ENVIRONMENTAL SCIENCES (0768)
 Abstract: Cloud-climate interannual variations associated with the
           anomalous SST over the tropical region are investigated
           by using five-year monthly mean SST, ERBE data and
           CCM2/AMIP simulation. Cloud effects, in terms of cloud-
           radiative forcing (CRF), on the atmospheric responses to
           the tropical SST anomaly, associated with El Nino and the
           Southern Oscillation (ENSO) event, have been studied by
           using the NCAR CCM2.
           The convection anomalies in terms of anomalous CRF appear
           to be in east-west anti-phase dipole patterns with
           ascending (descending) vertical motion over the central
           Pacific Ocean and descending (ascending) vertical motion
           over the maritime continent-South Pacific Convergence
           Zone (SPCZ) and Brazil for the positive (negative)
           anomalous SST presented over the central-eastern Pacific
           Ocean. It is suggested that the opposite type of strong
           anomalous SST may produce a reverse Walker circulation in
           the anomalous sense, which could be interpreted as a
           In our simulation, cloud effects intensify the model
           Southern Oscillation response to the SST anomaly with a
           negative SLP anomaly in the central-eastern Pacific Ocean
           and a positive anomaly in the western Pacific-Indian
           Ocean. Cloud effects produce a warming (cooling)
           throughout the troposphere and a cooling (warming) at the
           tropopause in the central-eastern Pacific (western
           Pacific-Indian Ocean). The zonal wind change associated
           with the temperature change shows a enhanced model
           Southern Oscillation with an westerly (easterly) anomaly
           at low levels and a easterly (westerly) anomaly at high
           levels in the central-eastern Pacific (western Pacific-
           Indian Ocean). The cloud effect weakens the intensity of
           the normal Walker circulation for both the east and west
           The longwave heating rate change due to cloud amount
           change, which is associated with the SST anomaly,
           presents a warming (cooling) at low levels and a cooling
           (warming) at high levels in the east (west). This change
           destabilizes (stabilizes) the troposphere and enhances
           (suppresses) deep convection and latent heat release in
           the east (west). The shortwave heating rate change is
           opposite to the longwave heating rate change and plays an
           opposite role to the cloud longwave heating as it
           interacts with deep convection and latent heat release.
           Cloud has a large impact on the latent heating response
           to the SST anomaly with a strong warming (cooling)
           presented in the middle-upper troposphere in the east
           (west), i.e., cloud effects enhance (suppress) the deep
           convection in the east (west) with the ENSO type SST
           anomaly over the central-eastern Pacific.
    Title: Frictional convergence and the Madden-Julian oscillation
   Pub No: 9983516
   Author: Maloney, Eric Daniel
   Degree: PhD
     Date: 2000
    Pages: 138
  Adviser: Hartmann, Dennis L.
     ISBN: 0-599-89556-X
   Source: DAI-B 61/08, p. 4213, Feb 2001
 Abstract: Frictional surface convergence is shown to be important
           to the Madden-Julian oscillation (MJO) of the tropical
           troposphere using both observations and an atmospheric
           general circulation model (GCM). An observed composite
           MJO lifecycle is created using the first two empirical
           orthogonal functions of the NCEP 850 mb equatorial zonal
           wind as an index. Lower tropospheric water vapor over the
           Indian and western Pacific Oceans is significantly
           correlated with 1000 mb convergence during an MJO
           lifecycle. Frictional convergence may help to
           precondition the atmosphere for strong MJO convection by
           moistening the lower troposphere.
           Several convection schemes were substituted into the NCAR
           CCM3 GCM to improve model intraseasonal variability. The
           McRAS convection scheme, that includes a parameterization
           of unsaturated convective downdrafts and a relative
           humidity threshold for convective initiation, produces
           much improved intraseasonal variability over the default
           CCM3 convection scheme of Zhang and McFarlane.
           Intraseasonal variability is decreased for higher
           relative humidity thresholds with McRAS, contrary to the
           results of Wang and Schlesinger (1999). Removal of the
           unsaturated convective downdraft and rain evaporation
           parameterization greatly diminishes intraseasonal
           variability in zonal winds and precipitation. An
           excessively dry tropical troposphere is most likely
           responsible for the degradation of intraseasonal
           Low-level MJO specific humidity anomalies in the model
           are strongly correlated with surface convergence. This
           strong coupling with surface convergence causes model MJO
           convection anomalies to be associated with 850 mb
           easterly perturbations. Where the eastward propagating
           model MJO signal is strong, low-level specific humidity
           anomalies precede those at middle and upper levels.
           Surface convergence affects model convection by
           moistening the lower troposphere, suggesting that surface
           convergence during an MJO lifecycle may help to
           precondition the atmosphere for intense MJO convection.
           Meridional convergence onto the equator contributes most
           of the anomalous convergence during an MJO lifecycle in
           both the model and observations. The wind-induced surface
           heat exchange (WISHE) mechanism of Emanuel (1987) and
           Neelin and Yu (1994) does not significantly contribute to
           the MJO-like mode in the CCM3 with McRAS.
    Title: Sensitivity studies for incorporating the direct effect
           of sulfate aerosols into climate models
   Pub No: 9965865
   Author: Miller, Mary Rawlings Lamberton
   Degree: PhD
     Date: 2000
    Pages: 271
  Adviser: Dickinson, Robert E.
     ISBN: 0-599-70484-5
   Source: DAI-B 61/03, p. 1455, Sep 2000
 Abstract: Aerosols have been identified as a major element of the
           climate system known to scatter and absorb solar and
           infrared radiation, but the development of procedures for
           representing them is still rudimentary. This study
           addresses the need to improve the treatment of sulfate
           aerosols in climate models by investigating how sensitive
           radiative particles are to varying specific sulfate
           aerosol properties. The degree to which sulfate particles
           absorb or scatter radiation, termed the direct effect,
           varies with the size distribution of particles, the
           aerosol mass density, the aerosol refractive indices, the
           relative humidity and the concentration of the aerosol.
           This study develops 504 case studies of altering sulfate
           aerosol chemistry, size distributions, refractive indices
           and densities at various ambient relative humidity
           conditions. Ammonium sulfate and sulfuric acid aerosols
           are studied with seven distinct size distributions at a
           given mode radius with three corresponding standard
           deviations implemented from field measurements. These
           test cases are evaluated for increasing relative
           humidity. As the relative humidity increases, the complex
           index of refraction and the mode radius for each
           distribution correspondingly change. Mie theory is
           employed to obtain the radiative properties for each case
           study. The case studies are then incorporated into a box
           model, the National Center of Atmospheric Research's
           (NCAR) column radiation model (CRM), and NCAR's community
           climate model version 3 (CCM3) to determine how sensitive
           the radiative properties and potential climatic effects
           are to altering sulfate properties.
           This study found the spatial variability of the sulfate
           aerosol leads to regional areas of intense aerosol
           forcing (W/m<super>2</super>). These areas are
           particularly sensitive to altering sulfate properties.
           Changes in the sulfate lognormal distribution standard
           deviation can lead to substantial regional differences in
           the annual aerosol forcing greater than 2 W/m<super>
           2</super>. Changes in the aerosol chemical composition
           can lead to regional changes in the aerosol forcing
           greater than 0.5 W/m<super>2</super>. The relative
           humidity is shown to greatly influence the aerosol
           optical properties. Given the differences in aerosol
           forcing found due to varying sulfate properties, this
           study does not encourage the use of a single aerosol
           distribution to represent sulfate particles of all air
    Title: Estimate of the climate impact of biomass burning
           aerosols in the Amazon by using NCAR-CCM3
   Pub No: 9927482
   Author: Montero Martinez, Martin Jose
   Degree: PhD
     Date: 1999
    Pages: 148
  Adviser: Dickinson, Robert E.
     ISBN: 0-599-27189-2
   Source: DAI-B 60/04, p. 1656, Oct 1999
 Abstract: It has been recognized that pollution (sulfate aerosols),
           and biomass burning (smoke aerosols), are the two
           principal anthropogenic sources that are now influencing
           the global climate. Researchers are currently trying to
           provide better estimates of the climate impact due to
           anthropogenic aerosols and to lower the range of
           uncertainty in their calculations. Estimations suggest a
           global-mean direct radiative forcing in the range from
           &ndash;0.3 to &ndash;1.0 W/m<super>2</super> for biomass
           burning aerosols, with a similar value for the indirect
           effect. The primary goal of this work is to estimate the
           global climate impact due to smoke aerosols from biomass
           burning using the National Center for Atmospheric
           Research (NCAR)-Community Climate Model, Version 3
           (CCM3). Analysis of data from the AERONET project was
           used to define a more realistic average of the
           optical/radiative properties of smoke particles from
           biomass burning in the Amazon, and also, to provide
           relevant information about the cloud-aerosol
           interactions. This information was put into the context
           of the NCAR-CCM3 BATS, and then the cloud/radiative
           scheme of the standard model was modified to include both
           the direct and indirect effects of the aerosols. Our
           results suggest an annual global mean &ldquo;total&rdquo;
           (direct+indirect) radiative forcing of about &ndash;0.08
           W/m<super>2</super>. This value is in good agreement with
           recent estimations considering that we are only including
           the biomass burning aerosols in the Amazon region. In
           addition, our model simulations provide a
           &ldquo;total&rdquo; radiative forcing of about &ndash;10
           W/m<super>2</super> for the Amazon region during the dry-
           season, where the indirect effect is responsible for
           about 80% of this value. This research hopefully will
           contribute to the aerosol climate modeling area by
           providing a scheme in which both the direct and indirect
           effects of the aerosol can be explicitly represented in a
    Title: Jet formation and equatorial superrotation in Jupiter's
           atmosphere: Numerical modelling using a new efficient
           parallel code
   Pub No: 3037750
   Author: Rivier, Leonard Gilles
   Degree: PhD
     Date: 2001
    Pages: 103
  Adviser: Polvani, Lorenzo M.
     ISBN: 0-493-50951-8
   Source: DAI-B 62/12, p. 5770, Jun 2002
           AND ASTROPHYSICS (0606)
 Abstract: Using an efficient parallel code solving the primitive
           equations of atmospheric dynamics, the jet structure of a
           Jupiter like atmosphere is modeled.
           In the first part of this thesis, a parallel spectral
           code solving both the shallow water equations and the
           multi-level primitive equations of atmospheric dynamics
           is built. The implementation of this code called BOB is
           done so that it runs effectively on an inexpensive
           cluster of workstations. A one dimensional decomposition
           and transposition method insuring load balancing among
           processes is used. The Legendre transform is cache-
           blocked. A &ldquo;compute on the fly&rdquo; of the
           Legendre polynomials used in the spectral method produces
           a lower memory footprint and enables high resolution runs
           on relatively small memory machines. Performance studies
           are done using a cluster of workstations located at the
           National Center for Atmospheric Research (NCAR). BOB
           performances are compared to the parallel benchmark code
           PSTSWM and the dynamical core of NCAR's CCM3.6.6. In both
           cases, the comparison favors BOB.
           In the second part of this thesis, the primitive equation
           version of the code described in part I is used to study
           the formation of organized zonal jets and equatorial
           superrotation in a planetary atmosphere where the
           parameters are chosen to best model the upper atmosphere
           of Jupiter. Two levels are used in the vertical and only
           large scale forcing is present. The model is forced
           towards a baroclinically unstable flow, so that eddies
           are generated by baroclinic instability. We consider
           several types of forcing, acting on either the
           temperature or the momentum field. We show that only
           under very specific parametric conditions, zonally
           elongated structures form and persist resembling the jet
           structure observed near the cloud level top (1 bar) on
           Jupiter. We also study the effect of an equatorial heat
           source, meant to be a crude representation of the effect
           of the deep convective planetary interior onto the outer
           atmospheric layer. We show that such heat forcing is able
           to produce strong equatorial superrotating winds, one of
           the most striking feature of the Jovian circulation.
    Title: Precipitation simulation in global climate models: Impact
           of horizontal resolution and improved land surface scheme
   Pub No: 9626486
   Author: Shaikh, Muhammad Javed
   Degree: PhD
     Date: 1996
    Pages: 246
  Adviser: Dickinson, Robert E.
   Source: DAI-B 57/04, p. 2438, Oct 1996
 Abstract: This dissertation examines the improvement in the
           simulation of precipitation by using a high resolution
           model and a highly physically-based land surface scheme.
           To investigate the impact of the horizontal resolution on
           the simulated fields, six versions of NCAR's latest
           Community Climate Model (CCM2) are analyzed. The
           simulations were conducted for periods of 4-20 years. To
           study the effects of the improved land surface scheme in
           the model, the coupling of BATS to the standard model and
           a revised version of CCM2 with modified optical
           properties for clouds are compared.
           It is evident that the refinement of the model resolution
           adds further accuracy to the simulation. The increased
           resolution of the model improves the definition of the
           major mountain ranges and the average surface topography
           of lands and oceans. The taller mountains in a high
           resolution model play an important role in modifying the
           vapor flux over the land surface. They obstruct more
           water flow and reduce the specific humidity and the total
           precipitable water over land. The vapor transport to the
           land surface is also partly influenced by the changes in
           the large scale circulation. The adjusted surface
           pressures over the sea surface levels simulate more
           closely the position and magnitude of the observed
           pressure systems.
           A major problem in CCM2 is the excessive surface
           radiation, which is up to 50-100 Wm$/sp[-2]$ higher than
           the Surface Radiation Budget (SRB) dataset over vast
           areas in the summer hemisphere. At a high resolution, the
           lower land clouds reduce the planetary albedo and allow
           more solar radiation to reach the surface. But in spite
           of this, the high resolution model improves the amount
           and distribution of the land precipitation. The RCCM2-
           BATS model, which increases the optical thickness of the
           summer clouds, substantially reduces the overestimate of
           the surface radiation. This helps reduce the large
           discrepancies obtained in summer precipitation. The land
           surface scheme predicts the regional climate more
           accurately than the standard model, but its performance
           is restricted by the inputs and does not show any major
           improvement in precipitation predictions which are mainly
           controlled by various processes in the atmospheric model.
    Title: Snow fractional cover in land surface schemes (Alaska)
   Pub No: 3113143
   Author: Slater, Andrew G.
   Degree: PhD
     Date: 2003
    Pages: 186
  Adviser: Lynch, Amanda H.
     ISBN: 0-496-60639-9
   Source: DAI-B 64/11, p. 5569, May 2004
 Abstract: Land surface scheme forcing data sets are constructed
           using observations from the Seward Peninsula in Alaska.
           Non-parametric statistical techniques are successfully
           applied to fill data gaps in precipitation.
           An extensive validation of the NCAR models, LSM and CLM,
           is undertaken using a variety of independent data such as
           turbulent fluxes measured via eddy correlation
           techniques. The models perform consistently compared to
           pervious studies, that is, CLM is a warmer dryer model
           than LSM. Several weaknesses regarding parameter
           assignment in the summer albedo, snow density and the
           lack of organic soil layers are identified in CLM. The
           additional physics in CLM are shown to perform better
           than the simpler LSM if the appropriate parameters are
           Using synthetic data generated from observations, a
           spatially implicit comparison between the models and
           remotely sensed products is made. The aim is to assess
           the ability of the models to capture the albedo and
           fractional snow cover over a larger area than a point.
           The models cannot emulate the observed variability of the
           region, even that within vegetation classes. Part of this
           stems from the uniform view of vegetation as well as the
           ever problematic snow albedo parameterization. The MODIS
           and AVHRR albedo products are compared against each other
           and can demonstrate consistent performance suggesting
           they may provide a key to the fractional cover issue.
           The implicit inclusion of a partial snow cover in the
           models is compared to explicit simulation of snow covered
           and snow free areas. Some differences are evident in the
           resulting mean surface temperature. The impact of non-
           linear behaviour in surface temperature with fractional
           cover was investigated through an upscaling exercise.
           Results suggest the models are largely scale invariant in
           off-line simulations.
    Title: Carbonate diagenesis and paleoclimate of the Desmoinesian
           Lower Ismay zone, Paradox Basin, southeast Utah
   Pub No: 3113055
   Author: Smith, Linda Sue Susanna
   Degree: PhD
     Date: 2003
    Pages: 251
  Adviser: Mitterer, Richard M.
     ISBN: 0-496-60551-1
   Source: DAI-B 65/06, p. 2821, Dec 2004
  Subject: GEOLOGY (0372); PALEOECOLOGY (0426); PHYSICAL
           OCEANOGRAPHY (0415)
 Abstract: The Paradox Basin is located in present day four corner
           area of SE Utah, USA. However, during the Desmoinesian
           the basin migrated from the equator to 5&deg;N.
           Carbonates developed on topographic highs during late
           transgressive system tracts (TST) and highstand system
           tracts (HST). Lowstand system tracts (LST) are
           characterized by more than 33 evaporite cycles. Platform
           carbonates and basin evaporite sequences are tied by
           cyclic regional dolomudstones that blanket both carbonate
           and evaporite sediments in the early TST. Detail analyses
           of the Lower Ismay Zone reveals rapid sea level falls
           within system tracts and climate fluctuations from
           extreme aridity to monsoon periods.
           Paleoclimate interpretation was enhanced by applying
           geochemical and petrographic analyses of all three
           depositional systems carbonate, evaporite and
           dolomudstone) of the Lower Ismay and Evaporite Cycle 3.
           Paragenetic sequences were constructed from four
           different horizons within the Lower Ismay Zone. These
           sequences identified three rapid sea level falls within
           the Lower Ismay. Meteoric cements in the lowest horizon
           indicate wetter periods in contrast to the more arid
           upper horizons that show well-developed sabkha fabrics
           and evaporite cements.
           To better understand the effect climate change played on
           sea level fluctuations and diagenesis, climate models
           were run in the NCAR CSM3 program. The Late Carboniferous
           Earth plate configuration is the geographic base to test
           two obliquity extremes (21.5&deg; and 24.5&deg;). Results
           from these simulations suggest a significant shift in the
           Inter-tropical Convergence Zone (ITCZ) for 21.5&deg; and
           24.5&deg; obliquities. This shift indicates that wet
           sediment cycles in the Desmoinesian of Wyoming are out of
           phase with those in the Paradox Basin. A strong seasonal
           shift in wind direction with cyclonic movement occurs in
           both models in the Paleo-Tethys, suggesting the strong
           monsoon climate in this region is not driven by
           obliquity. The maximum solar insolation latitudinal band
           is expanded in the January/July 24.5&deg; obliquity
           model. Also, the tropical sea surface temperatures show a
           broader warm band at the equator in high obliquity
    Title: Examining vegetation and climate interactions:  A
           comparative study of different vegetation types and their
           impacts on atmospheric energetics and dynamics at the
           regional and global scales
   Pub No: 3128110
   Author: Snyder, Peter K.
   Degree: PhD
     Date: 2004
    Pages: 205
  Adviser: Foley, Jonathan A.
     ISBN: 0-496-75373-1
   Source: DAI-B 65/04, p. 1914, Oct 2004
           SCIENCES (0768)
 Abstract: It is widely recognized that terrestrial ecosystems play
           an important role in maintaining the climate system.
           Vegetation influences the climate through complex
           biophysical exchanges of energy, water, and momentum, and
           changes to these fluxes can have a significant impact on
           the climate. As a result, natural and anthropogenic land
           use and land cover change can have a substantial impact
           on climate.
           Using a coupled atmosphere-biosphere model (CCM3-IBIS),
           the influence of vegetation on climate at the regional,
           continental, and global scales is evaluated. Comparison
           of the vegetation removal and control simulations
           illustrates the climatic influence of particular
           vegetation types. The mechanisms responsible for
           influencing the climate away from the surface, into the
           atmosphere, and to remote regions not directly coupled to
           the surface forcing are investigated.
           It was found that vegetation biomes from around the globe
           have a significant influence on the climate system with
           the response dictated by the vegetation type, the
           geographic location, and the physical processes important
           for that biome. For instance, the climate response to
           tropical deforestation is a surface warming due to a
           large reduction in latent cooling. In contrast, boreal
           deforestation causes a large surface temperature cooling
           due to a higher surface albedo.
           It was also shown that changes in the land surface
           resulting from land cover change affect the transport of
           energy throughout the three-dimensional atmosphere and
           can change the climate in adjacent regions or locations
           far removed from the surface forcing through modification
           of deep moist convection.
           Finally, the results indicate that tropical deforestation
           can have a significant influence on the northern
           hemisphere climate through changes to the atmospheric
           thermodynamics and general circulation. Tropical
           deforestation was found to influence the northern
           hemisphere climate through changes to the high-level
           tropical outflow, the strength and structure of the East
           Asian Jet, the poleward shift in the high-latitude jet,
           and changes to the Eurasian climate through amplification
           of the positive phase of the North Atlantic Oscillation.
    Title: A numerical modeling study of the coupled variability of
           Lake Victoria in eastern Africa and the regional climate
   Pub No: 3033715
   Author: Song, Yi
   Degree: PhD
     Date: 2000
    Pages: 184
  Adviser: Xie, Lian; Semazzi, Fredrick H. M.
     ISBN: 0-493-46685-1
   Source: DAI-B 62/12, p. 5770, Jun 2002
 Abstract: The objective of this investigation was to investigate
           and study the coupled atmosphere-lake climate system over
           the Lake Victoria basin, and determine the corresponding
           physical mechanisms that are involved. The primary
           research vehicle for the investigation is a fully coupled
           model of the regional climate of Eastern Africa and Lake
           Victoria which has been developed and applied in this
           study. The atmospheric component of the model is the NCAR
           Regional Climate Model (RegCM2).
           The results show that the bythemetry and geometry of the
           lake play a fundamental role in determining the
           climatology of Lake Victoria. There exists Kelvin-like
           waves in the thermocline trapped along the coast and they
           propagate clockwise around Lake Victoria with periodicity
           of about 30 days. Preliminary comparison of the coupled
           RegCM2-POM model simulation results with the observations
           indicates that the model produces more realistic lake
           surface temperatures (LST) and rainfall over and around
           the lake than the standard version of RegCM2 in which a
           simple one dimensional thermal diffusion lake model is
           Over Eastern Africa, the regional climate variability is
           significantly influenced by the circulation over the Lake
           Victoria basin. The interaction between the lake-land
           breeze and the prevailing northeasterly flow accounts for
           the asymmetry in the distribution of the diurnal rainfall
           variations and the southwestward movement of the dominant
           bands of divergence/convergence.
           During the 1982 El Nino when the averaged LST over the
           lake was higher than that during the normal year, the LST
           gradient was weakened along the SW-NE axis over the lake
           by the strong lake circulation. The hydrodynamics of the
           lake play an important role in determining the coupled
           variability of the lake circulation and the lake basin-
           wide climatic conditions.
           The potential climate change resulting from total
           clearing of the tropical rain forests in Africa was also
           investigated by the standard version of the NCAR CCM3
           global climate model. Over Eastern and Western Africa the
           impact of deforestation is primarily characterized by
           reduction in rainfall, however the CCM3 resolution of T42
           which we have adopted may not be adequate to resolve the
           large contrasts in terrain and vegetation types. A
           striking result is that the strong remote response of the
           Southern Africa region to deforestation over Central
           Africa. (Abstract shortened by UMI.)
    Title: The influence of atmospheric chemistry and climate on
           atmosphere-biosphere interactions
   Pub No: 3095758
   Author: Steiner, Allison L.
   Degree: PhD
     Date: 2003
    Pages: 133
  Adviser: Chameides, William L.
     ISBN: 0-496-43383-0
   Source: DAI-B 64/06, p. 2718, Dec 2003
 Abstract: A modeling tool is developed to investigate the feedbacks
           between the terrestrial biosphere and the atmosphere, two
           major components of the Earth system. Changes in
           environmental variables such as solar radiation,
           temperature, and atmospheric water vapor can influence
           the water and energy balances at the land surface. These
           changes can, in turn, strongly impact vegetation
           processes, including transpiration, photosynthesis, and
           the emission of biogenic volatile organic compounds
           (VOCs). These processes are part of the complex
           mechanisms within the Earth system that can feedback upon
           each other and create unexpected responses.
           A biogenic VOC emissions algorithm is developed for use
           within the modeling system. Biogenic VOC emissions,
           including isoprene, monoterpenes and ORVOCs, account for
           over half of the total amount of VOCs emitted in East
           Asia. Additionally, the emissions can be significantly
           affected by both land cover changes and rising
           CO<sub>2</sub> concentrations. Because of their
           importance in predicting ground-level ozone, these
           terrestrial biosphere-atmosphere feedbacks could have an
           important influence on air quality prediction and
           As part of the coupled climate-land surface model
           development, a new land surface model, the Common Land
           Model (CLM0) is integrated into a regional climate model
           (RegCM2). For a yearlong simulation over East Asia, the
           RegCM2/CLM0 slightly improves the winter cold bias that
           is present in the former RegCM2 simulation and reproduces
           the seasonal cycle of climate in the region.
           Three-dimensional fields of anthropogenic aerosols are
           added to the RegCM2/CLM0 to investigate biosphere-
           atmosphere feedbacks. The presence of aerosols reduces
           leaf temperatures crucial for photosynthesis and biogenic
           VOC emissions. In a five-day summertime simulation that
           includes the so-called direct effect of aerosols, there
           is a 5&ndash;8 degree reduction in leaf temperatures
           within the canopy. In the absence of aerosols, leaf
           temperatures are beyond the optimum leaf temperature,
           which slows the photosynthetic rate. The presence of
           aerosols reduces these temperatures below the optimum,
           leading to an increase in photosynthesis and
           transpiration. This result indicates a previously
           undiscovered mechanism between aerosols and the
           terrestrial biosphere, and this could have significant
           implications for future studies in aerosol-climate
           interactions and the terrestrial carbon balance.
    Title: A one-dimensional mixed-layer ocean model for use in
           three-dimensional climate simulations
   Pub No: 9830538
   Author: Stephens, Monica Yvette
   Degree: PhD
     Date: 1998
    Pages: 196
  Adviser: Maxey, Martin
     ISBN: 0-591-83340-9
   Source: DAI-B 59/04, p. 1689, Oct 1998
 Abstract: A study has been made of the dynamic interactions between
           the surface layer of the oceans and the atmosphere as it
           relates to global climate variations. The atmospheric
           conditions are simulated numerically with the NCAR
           Atmospheric General Circulation Model (AGCM), CCM3. A
           one-dimensional mixed layer ocean model (MLOM) for the
           upper ocean has been developed and coupled to the AGCM.
           The mixed-layer model simulates vertical ocean dynamics
           and demonstrates the effect of mixed-layer depth and
           convective instability on the sea surface temperature and
           on the formation of sea-ice. The purpose of the coupling
           was to make direct calculations of mixed-layer depth
           convection in a climate simulation using CCM3. The
           current mixed-layer model that is used in CCM3, known as
           the slab ocean model, contains seasonally and spatially
           specified depths. It also uses a prescribed horizontal
           heat flux (QFLUX) that determines the vertical heat
           transport and models the large scale heat transport by
           ocean currents. The MLOM allows the coupled simulation to
           capture the physics of sea-ice and mixed-layer
           interaction, and the effects of the seasonal variation of
           mixed-layer depth and entrainment on the sea surface
           temperature. This model has also been run in stand-alone
           simulations with data derived from the AGCM, CCM1. These
           simulations show that the MLOM is able to simulate SSTs
           and depths that closely approximate the observed climate.
           The addition of the variable-depth mixed-layer model to
           CCM3 provides several potentially interesting features to
           the climate simulation. The mixed-layer model deals
           explicitly with vertical heat transport--one of the
           phenomena parameterized by QFLUX in current slab ocean
           models, calculates mixed-layer depth which is currently
           seasonally and spatially specified in CCM3, and
           introduces a negative feedback for sea-ice growth from
           the interaction between the mixed layer and sea ice. An
           explicit calculation of mixed-layer depth and vertical
           heat flux will make CCM3 better able to make physically
           accurate predictions of sea surface temperature and sea
           ice for past and future climate scenarios since these
           quantities will no longer be tuned to meet present-day
           conditions. Currently, CCM3 coupled to the slab ocean
           model only includes the ice-albedo feedback mechanism
           which is a positive feedback for sea-ice growth.
           Including the negative feedback from mixed-layer and sea-
           ice interaction provides a means for studying the
           competing effects of both negative and positive feedbacks
           on sea-ice growth.
    Title: Tropospheric constituent variability associated with
           baroclinic waves
   Pub No: 9814700
   Author: Stone, Elizabeth Mary
   Degree: PhD
     Date: 1997
    Pages: 111
  Adviser: Stanford, John L.
     ISBN: 0-591-65736-8
   Source: DAI-B 58/11, p. 6018, May 1998
 Abstract: Baroclinic waves are fundamental to tropospheric
           dynamics, and therefore it is of interest to understand
           their impact on atmospheric constituents. Here we examine
           the dynamics of the troposphere during baroclinic wave
           life cycles and study in detail their influence on
           constituent transport. We incorporate two methods of
           studying the atmosphere. The first is the study of
           observations made by remote sensing of the atmosphere by
           satellites. The second is the examination of the
           atmosphere through the use of an atmospheric general
           circulation model.
           In an observational case study, we used upper
           tropospheric water vapor measurements from the Upper
           Atmosphere Research Satellite Microwave Limb Sounder to
           investigate the structure and evolution of eastward
           traveling medium-scale wave features in Southern
           Hemisphere summertime and found that the water vapor
           field is well correlated with meteorological fields and
           derived potential vorticity fields. These results are
           consistent with model paradigms for the structure and
           evolution of baroclinic disturbances.
           In order to study the details of transport associated
           with baroclinic waves, we build upon the successes of
           other studies which use general circulation models to
           simulate baroclinic wave life cycles. These nonlinear
           simulations have shown that the wave evolution consists
           of baroclinic growth, maturity and barotropic decay. In
           our study, two life cycles are simulated with the
           National Center for Atmospheric Research (NCAR) community
           climate model (CCM2), starting with baroclinically
           unstable initial conditions similar to those used by
           Thorncroft et al. (1993). The two life cycles differ in
           the strength and sense of the horizontal shear of the
           zonal wind. This strongly influences the behavior which
           ensues. In terms of potential vorticity-potential
           temperature diagnostics, the basic case is characterized
           by thinning troughs which are advected anti-cyclonically
           and equatorward, while the anomalous case has broadening
           troughs which wrap up cyclonically and poleward. In order
           to investigate transport during these two life cycles,
           four passive tracers are included in the simulation to be
           advected by the semi-Lagrangian transport scheme of CCM2.
           The resulting tracer budgets are analyzed in terms of the
           transformed Eulerian mean constituent transport
           formalism. Results show a net upward and poleward
           transport and a strong influence of the eddy flux term on
           the time tendency of the tracer. The largest transport
           occurs during the nonlinear growth stage of the life
           cycle. We also find that the transport varies little with
           the initial tracer distributions.
    Title: Diurnal variation of deep convection over the tropical
           western Pacific
   Pub No: 3115874
   Author: Sun, Moguo
   Degree: PhD
     Date: 2003
    Pages: 123
  Adviser: Cess, Robert D.
     ISBN: 0-496-63311-9
   Source: DAI-B 64/12, p. 5987, Jun 2004
 Abstract: An integration of observational and modeling studies of
           the diurnal variation of deep convection is presented in
           this thesis. The thesis consists of three parts:
           (1)&nbsp;an observational study on the diurnal variation
           of deep convection, (2)&nbsp;Adequacy of the National
           Center for Atmospheric Research (NCAR) Community Climate
           Model version 2 (CCM2) radiation scheme used in the
           modeling study and the sensitivity of outgoing longwave
           radiation (OLR) to atmospheric variables, (3)&nbsp;a
           modeling study on the mechanisms of the diurnal variation
           of deep convection.
           The observational study concentrates on the tropical
           western Pacific (TWP) during the Tropical Ocean Global
           Atmosphere-Coupled Ocean Atmosphere Response Experiment
           (TOGA-COARE) intensive observation period (IOP). Four
           months of Geostationary Meterorological Satellite (GMS)
           infrared (IR) temperature data are analyzed using IR
           temperature threshold and cloud cluster techniques. Both
           techniques reveal the same results, that deep convective
           cloudiness reaches a maximum in the early morning
           (03&ndash;06 local standard time (LST)). The feature is
           very different from lower level cloudiness that reaches a
           maximum in the late afternoon. A cross-plot of IR
           temperature versus time shows that the phase of the
           diurnal variation has an abrupt change around 220K and
           propagates downward. Spectral analysis is used to
           establish the robustness of the diurnal signal of deep
           Unlike previous studies using idealized simulations, this
           study employs a three-dimensional regional model, the
           NCAR Mesoscale Model Version 5 (MM5), to realistically
           simulate the diurnal variation. Before this, however, the
           study focuses on the radiation code of MM5, since
           radiation is particularly important in the modeling
           study. Comparisons between CCM2 radiation code used in
           MM5 and its updated CCM3 version are to show the adequacy
           of CCM2 radiation code in the current study. The
           comparisons show that both solar and longwave radiation
           produced by the CCM2 code are close to the later CCM3
           version under the clear sky condition. (Abstract
           shortened by UMI.)
    Title: Ensemble forecasting with the ensemble transform Kalman
   Pub No: 3147696
   Author: Wang, Xuguang
   Degree: PhD
     Date: 2004
    Pages: 75
  Adviser: Young, George S.
     ISBN: 0-496-06817-2
   Source: DAI-B 65/09, p. 4626, Mar 2005
 Abstract: A new initial perturbation generation method, the
           ensemble transform Kalman filter (ETKF), is introduced
           and compared with the breeding scheme. The ETKF generates
           initial perturbations by postmultiplying forecast
           perturbations by a transformation matrix. This matrix is
           chosen to solve the error covariance update equation for
           an optimal data assimilation scheme within the ensemble
           perturbation subspace. Version 3 of the community climate
           model (CCM3) developed at National Center for Atmospheric
           Research is used to test and compare the ETKF and
           breeding schemes. It is found that with only a little
           more computational expense, the ETKF samples initial
           condition uncertainties significantly better than the
           breeding. The ETKF ensemble mean and ensemble covariance
           are considerably more accurate than those of the
           A new method to center initial ensemble perturbations on
           the initial analysis is introduced and compared with the
           commonly used centering method of positive-negative
           paired perturbations. In the new method, called spherical
           simplex centering, one linearly dependent perturbation is
           added to a set of linearly independent initial
           perturbations to ensure that the sum of the new initial
           perturbations equals zero; the covariance calculated from
           the new initial perturbations is equal to the analysis
           error covariance estimated by the independent initial
           perturbations; and all the new initial perturbations are
           equally likely. 16-member CCM3 ETKF ensemble initially
           centered by the new method is found to be more skillful
           than that centered by the positive-negative paired
           A new ensemble postprocessing method that reduces
           seasonally averaged second moment errors of the ensemble
           forecasts is introduced. The method involves adding
           (&ldquo;dressing&rdquo;) independent sets of statistical
           perturbations to each member of a dynamical ensemble
           forecast. The new dressing method mathematically
           constrains the stochastic process used to generate the
           statistical perturbations so that it entirely removes
           seasonally averaged errors in the second moment measures.
           ETKF ensembles dressed with this new method are found to
           be more skillful than the undressed ETKF ensembles. It is
           shown that the previously proposed &ldquo;best
           member&rdquo; dressing method fails to reliably predict
           the second moment of the distribution of forecast errors
           whereas the new dressing method reliably predicts this
           second moment.
    Title: Spatial and temporal scales of precipitating tropical
           cloud systems
   Pub No: 3071048
   Author: Wilcox, Eric Martin
   Degree: PhD
     Date: 2002
    Pages: 131
  Adviser: Ramanathan, V.
     ISBN: 0-493-90927-3
   Source: DAI-B 63/11, p. 5293, May 2003
           REMOTE SENSING (0799)
 Abstract: Precipitation, radiative forcing, and aerosol scavenging
           in tropical cloud systems over the wintertime Indian
           Ocean are examined in satellite observations and global
           atmospheric simulations. Measurements of surface rain
           rate and top-of-atmosphere radiative fluxes from the TRMM
           satellite, as well as brightness temperature measurements
           from the METEOSAT-3 satellite, are used to identify the
           boundaries of cloud systems, track their evolution, and
           determine the spatial and temporal scales of cloud
           thermodynamic forcing. The resulting quantitative,
           statistical description of monsoonal cloud systems is
           compared with simulated cloud systems in the NCAR CCM3
           Monsoonal clouds span a spectrum of spatial scales from
           smaller than 25 km<super>2</super> to greater than
           10<super>7</super> km<super>2</super>. Atmospheric
           heating owing to precipitation and the cloud greenhouse
           effect, as well as surface cooling owing to cloud albedo,
           increases with the spatial scale of cloud systems. As a
           result, thermodynamic forcing of the monsoonal
           environment is dominated by the contribution from giant
           semi-permanent decks of overcast cloud that persist for
           days to weeks. Embedded within such cloud decks are
           numerous rain cells reaching up to 1 million square-
           kilometers because deep convection organizes into
           clusters of narrow overturning cells attached to a broad
           stratiform region of precipitation. A relatively few such
           mesoscale convective systems are greater than
           10<super>5</super> km<super>2</super>, yet are
           responsible for up to 70% of monsoonal precipitation. In
           contrast, simulated cloud systems in the model gently
           precipitate throughout their duration and everywhere
           within their boundaries. The model lacks a process that
           acts to organize convection into mesoscale episodic
           Precipitation is the principal means by which particulate
           pollution is removed from the atmosphere. The effect of
           model biases in the distribution of precipitation is
           tested by integrating satellite precipitation
           measurements into the MATCH chemical transport model.
           Mesoscale convective systems in the equatorial Indian
           Ocean are a substantial barrier to the transport of
           aerosols from South Asia to the Southern Hemisphere.
           Using observations of the spatial coverage of
           precipitation in the model reduces the amount of South
           Asian aerosol transported to the remote Northern
           Hemisphere by more than a factor of 2 compared to a
           simulation using model derived precipitation.
    Title: Biome-scale vegetation dynamics in North America since
           the last glacial maximum:  Maps and reconstructions from
           fossil pollen data and the testing of biogeography models
   Pub No: 9987861
   Author: Williams, John Warren
   Degree: PhD
     Date: 2000
    Pages: 266
  Adviser: Webb, Thompson III
     ISBN: 0-599-94261-4
   Source: DAI-B 61/09, p. 4630, Mar 2001
  Subject: PALEOECOLOGY (0426); PALYNOLOGY (0427)
 Abstract: The changing biome distributions reconstructed from
           fossil pollen data illustrate the dynamic nature of the
           vegetation in boreal and eastern North America since the
           last glacial maximum, 21,000 years ago. Biomes, defined
           as mixtures of plant functional types, emphasize
           structural changes in the vegetation in response to
           climate change. A primary use of these data-based biome
           maps is to test simulations produced by atmospheric
           general circulation models (AGCM's) and vegetation
           Biome maps for 6000 and 18,000 radiocarbon years B.P.
           were created as part of an international effort to
           assemble global paleovegetation maps for model testing.
           The biome reconstructions for 6000 years B.P. serve as a
           benchmark for a series of detailed evaluations of the
           biogeography models BIOME1 and BIOME3. A 2 x 2 factorial
           set of biome simulations was performed in which climate
           simulations for 6 rka from CCM1 and CCM3 each were used
           to run BIOME1 and BIOME3. Only the biome simulation
           produced by CCM1 and BIOME3 matched the data well.
           Sensitivity analyses revealed the poor performance of the
           CCM3/BIOME3 simulation primarily was caused by BIOME3
           being overly sensitive to secondary influences such as
           cloud cover and soil characteristics. This sensitivity
           builds a greater uncertainty into the BIOME3 simulations,
           particularly because cloudiness may not be accurately
           simulated by AGCM's.
           Biome distributions and individual pollen taxon
           abundances were mapped at 1000-year intervals to explore
           vegetation dynamics in eastern North America. Pollen
           assemblages without any close modern analog prevailed
           adjacent to the Laurentide Ice Sheet between 17,000 and
           12,000 kcal. These &ldquo;no-analog&rdquo; pollen samples
           indicate that late-glacial plant associations were unlike
           any today. Distinguished by the intermixing of boreal
           taxa with temperate taxa and herbs, these no-analog
           communities were likely a mixed parkland. CCM1
           simulations suggest that the no-analog vegetation grew
           under highly seasonal and dry climates. The temporal and
           spatial distribution of the no-analog vegetation strongly
           corresponds to the distribution of no-analog climatic
           conditions simulated by CCM1. The fastest rates of
           climatic and vegetation change occurred after the peak
           periods of no-analogs. Together, these lines of evidence
           support the hypothesis that vegetation is in equilibrium
           with climate at millennial timescales.
    Title: Optimal climate change signal detection using space-time
           empirical orthogonal functions
   Pub No: 9969031
   Author: Wu, Qigang
   Degree: PhD
     Date: 2000
    Pages: 116
  Adviser: North, Gerald R.; Kim, Kwang-Yul
     ISBN: 0-599-73836-7
   Source: DAI-B 61/04, p. 1998, Oct 2000
 Abstract: Space-time optimal filtering is used to estimate the
           response of the Earth's surface temperature to both
           natural and anthropogenic climate forcings over the past
           century. In this study, a variety of site/record-length
           configurations is used to represent the climate change
           signals, natural variability and observational data
           streams. The hypothesized space-time patterns of the
           response to the climate forcings are generated from an
           energy balance climate model (EBCM) and four coupled
           ocean/atmosphere general circulation models (GCMs). The
           natural variability statistics are estimated from 1000-
           year control runs from four coupled GCMs and then such
           natural variability is represented precisely in terms of
           the space-time empirical orthogonal functions (EOFs),
           obtained directly from the lagged space-time covariance
           matrix on the approximate time interval. Optimal filters
           are constructed using the hypothesized space-time
           patterns along with the EOF-represented natural
           variability. The constructed filters are then applied to
           the observational surface temperature data. With a
           proposed orthogonalization process, signals &lsquo;not-
           of-interest&rsquo; are optimally removed from the data
           stream in the estimation process.
           Comparisons show that the responses to the anthropogenic
           climate forcings generated from the EBCM and four GCMs
           are very similar to each other. A robust greenhouse-gas
           signal is detected in the observational surface
           temperature data, but with a weak amplitude only about
           60&ndash;70% of that expected from the hypothesized
           signal patterns. An anthropogenic aerosol signal is very
           weak and not statistically significant. The estimated
           amplitude of aerosol indicates that four GCMs in this
           study might overpredict the response to greenhouse gas-
           plus-aerosol forcing by a factor of two to five. A
           volcanic signal is also robust but with about 65%
           expected amplitude. A solar-cycle signal is significant
           at a level of 90% and somewhat stronger than the
           amplitude of the EBCM-predicted response to the 11-year
           component of the solar luminosity variations.
           A simplified version of NCAR Community Climate Model
           (CCM2) is used to generate natural variability and an
           artificial solar signal for the optimal filtering scheme.
           A theoretical examination of whether detection
           performance could be enhanced by the addition of
           vertically distributed spatial information rather than
           information only on the surface is presented.
    Title: Single-column modeling: Methodology and application to
           the evaluation of cumulus convection schemes in GCMs
   Pub No: 9927551
   Author: Xie, Shaocheng
   Degree: PhD
     Date: 1998
    Pages: 126
  Adviser: Zhang, Minghua
     ISBN: 0-599-27273-2
   Source: DAI-B 60/04, p. 1657, Oct 1999
 Abstract: I have studied two subjects on single-column modeling in
           this thesis: One is related with interfacing single-
           column models (SCMs) with observations. The other is
           related to the causes of large errors in SCM simulations.
           Problems with the discrepancy in the time resolution
           between observations and SCMs were studied through
           modeling sensitivity tests. These include study of the
           model sensitivity to time steps and to the time-averaged
           large-scale forcing. Experiments show that most of the
           simulation results are sensitive to time steps but
           insensitive to the use of time-averaged large-scale
           forcing within the 3-hour time period (the typical
           sampling frequency of observations). The SCM, along with
           a 3-hour time step, produces serious errors and causes
           computational instability. Therefore, a time step, that
           is the same as used in GCMs, is the most optimal time
           step for SCMs.
           Causes of large errors in SCM simulations were
           investigated by focusing on deficiencies in the model
           cumulus parameterizations. Three cumulus convection
           schemes were implemented in the NCAR CCM3 Single-Column
           Model and were evaluated by the ARM 1995 summer Intensive
           Observing Period (IOP) data. They are the CCM3 deep
           convection scheme, the moist convective adjustment (MCA)
           scheme, and the KUO scheme. The simulation results show
           that the SCM with the CCM3 deep convection scheme
           produces large warm and dry biases, while the SCMs with
           the other two convection schemes produce large cold and
           wet biases.
           The triggering conditions of convection in these three
           schemes have been found to account for a significant part
           of these biases. Based on the observations, a
           modification is made to the triggering condition of the
           CCM3 deep convection scheme to account for the dynamical
           influences of circulation on the initiation of
           convection. The modified scheme is tested using the ARM
           data. Improved simulation results were obtained.
           Modifications of triggering conditions were also made to
           the MCA and KUO schemes to illustrate possible
           improvements to SCM simulations.
           Additional tests were made to the CCM3 deep convection
           scheme by using the ARM 1997 summer IOP and the GATE
           phase III data. These tests confirm the results obtained
           from the ARM 1995 experiments.
    Title: Intraseasonal oscillation in an idealized general
           circulation model
   Pub No: 9634876
   Author: Yi, Yuhong
   Degree: PhD
     Date: 1996
    Pages: 94
  Adviser: North, Gerald R.
   Source: DAI-B 57/06, p. 3796, Dec 1996
 Abstract: The intraseasonal oscillations (ISOs) that appear in the
           perpetual equinox simulation of a spectral general
           circulation model (GCM) have been systematically studied.
           The NCAR Community Climate Model (CCM2, R15) has been
           implemented on an idealized planet with no ocean and no
           topography and with uniform boundary conditions, such as
           constant vegetation and surface albedo, and also has been
           integrated for 15 years.
           The wave-frequency distributions at different latitudes,
           obtained from space-time spectral analysis, show that
           well-defined spectral peaks occur in the eastward
           propagating wave components of the zonal and meridional
           winds, especially in the wavenumber one component which
           has a period of about 40 days. Wavenumber 2 travels at a
           slightly slower phase speed and takes about 50-54 days to
           encircle the globe once.
           A phase-shifting composite procedure has been used to
           further demonstrate the three-dimensional spatial
           structure of the model's ISO. The horizontal structure of
           the simulated oscillation is dominated by the eastward-
           moving wavenumber 1; wavenumber 2, however, is also
           important and its amplitude is about half that of
           wavenumber 1. ISOs exhibit a baroclinic structure in the
           tropics but a barotropical structure in the extratropics.
           These results suggest that ISOs are internally generated
           and maintained by the model dynamics and physics.
           A newly-developed time-varying spectral analysis has been
           used to study the characteristics of the temporal
           evolution of the intraseasonal oscillation, and thus to
           explore the mechanism of ISO maintenance. The most
           interesting and striking result is that the internally
           generated intraseasonal oscillation exhibits an episodic
           feature, referred to as intermittency of ISO. Large- and
           small-amplitude oscillations occur intermittently.
           Although intermittency is a frequently observed feature
           in nonlinear and turbulent flows, it appears to be the
           first time that such intermittency has heen documented in
           such model studies of the intraseasonal oscillation.
           The relationship between the zonal-mean flow and low-
           frequency waves has been investigated by using the EOF
           analysis. The results show that intense ISOs are
           associated with weak, broad and poleward-shifted
           subtropical jets as well as intense westerlies in the
           upper tropics; less intense ISOs are associated with
           strong, narrow and equatorward-shifted subtropical jets.
           This finding suggests that nonlinear interactions between
           the zonal-mean flow and low-frequency waves may be
           important for maintaining the episodic feature of the
           intraseasonal oscillations.
    Title: Validation of the coupled NCEP mesoscale spectral model
           and an advanced land surface model over the Hawaiian
   Pub No: 3151140
   Author: Zhang, Yongxin
   Degree: PhD
     Date: 2004
    Pages: 207
  Adviser: Cheng, Yi-Leng
     ISBN: 0-496-11098-5
   Source: DAI-B 65/10, p. 5195, Apr 2005
           SCIENCES (0768)
 Abstract: The National Centers for Environmental Prediction (NCEP)
           Mesoscale Spectral Model (MSM) coupled with an advanced
           Land Surface Model (LSM) is used in assessing the impact
           of the improved representation of island terrain and
           surface boundary conditions on simulating orographic
           effects and local circulations under various large-scale
           The surface variables predicted by the operational 10-km
           Regional Spectral Model (RSM) during a one-month period
           of 20 May through 20 June 2002 agree well with
           observations over the buoy stations in the Hawaiian
           waters. Over land with adequate representation of the
           terrain, the 1.5-km MSM provides better forecasts of
           surface variables than the 10-km RSM. Further
           improvements are achieved by coupling the MSM with the
           LSM. In particular, over-estimation of the surface wind
           speed and daytime cold biases in the MSM are largely
           corrected in the coupled MSM/LSM. The observed composite
           diurnal cycles of surface variables are also better
           forecasted by the MSM/LSM than the MSM.
           Evaluations of the 3-km MSM/LSM simulations are performed
           for sea breeze cases during 23 June to 28 June 1978 over
           northwest Hawaii. Except for 27 June, the model predicted
           onset time, duration and vertical extent of the sea
           breezes agree with observations. Sensitivity tests using
           the MSM/LSM demonstrate the non-trivial effects of
           surface properties and initialization processes on sea
           breeze behavior.
           Heavy rainfall and high wind events over the Hawaiian
           Islands display a large variability in rainfall and wind
           distribution related to complex terrain and local winds.
           The 10-km RSM/LSM reasonably resolves regional-scale
           weather features associated with significant synoptic
           systems but fails to accurately reproduce rainfall
           distribution and rate, and orographically amplified
           strong winds. The MSM/LSM, with its higher resolution
           (&le;3 km), has better capability in simulating localized
           rainfall distributions and airflows associated with these
           heavy rainfall and high wind events.
           Major model bias is that the MSM/LSM produces excessive
           rainfall on the windward side of steep mountains with
           little rainfall downstream of the mountain ranges.
    Title: An algorithm for retrieving surface downwelling longwave
           radiation:  A study of interactive physical mechanisms
   Pub No: 9958706
   Author: Zhou, Yaping
   Degree: PhD
     Date: 1999
    Pages: 131
  Adviser: Cess, Robert D.
     ISBN: 0-599-62066-8
   Source: DAI-B 61/01, p. 348, Jul 2000
 Abstract: In this thesis, a new algorithm for retrieving the
           surface downwelling longwave flux is developed based on a
           detailed study of radiation models and observational
           data. The radiation models used in this study, the Column
           Radiation Models (CRM) from the National Center for
           Atmospheric Research (NCAR) community climate model (CCM)
           version2 and version3 and the Moderate Resolution
           Transmittance (MODTRAN3) Code, are validated with the
           CAGEX (CERES/ARM/GEWEX) project version 1.1.2 data taken
           at the Atmospheric Radiation Measurement (ARM) Program
           Oklahoma Central Facility. Results show that the accuracy
           of the radiation model is quite consistent with the
           models' level of complexity for clear skies. For cloudy
           skies, the cloud input parameters from various
           instruments need careful examination and preprocessing.
           The discrepancy between model calculations and
           observations can be significantly reduced by choice of
           input parameters and by tuning the optical properties
           within the models.
           Detailed sensitivity tests are conducted on the CCM3CRM
           to study the effect of atmospheric temperature and water
           vapor profiles upon the clear sky surface and top of
           atmosphere outgoing longwave fluxes. The study shows that
           the surface downwelling longwave flux can be largely
           determined from only two parameters: the surface
           upwelling longwave flux and the total precipitable water
           vapor. Cloudy sky sensitivities are conducted with both
           CCM3CRM and Modtran3. Both models find the cloud base
           height to be the most important factor determining the
           surface downwelling longwave, especially for low clouds.
           However, when considering partial cloud cases in the real
           world, column cloud liquid water seems to be a better
           parameter for the cloudy sky algorithm.
           The ARM observations at the Oklahoma Central Facility and
           the Tropical Western Pacific (TWP) Manus Island are used
           in deriving and validating the algorithm. The
           observations show similar relations found in the
           sensitivity tests for both clear skies and all skies. An
           algorithm is derived based on 6 Intensive Observational
           Period (IOP) data set at the Oklahoma Central Facility
           The rms error of this algorithm ranges from 5.8% for the
           data taken at the same location and 2.5% for the TWP data
           The generality of this algorithm is further demonstrated
           with CCM3 simulation data. A single algorithm can be
           applied to six geographically different areas, with rms
           errors ranging from 1.4% to 5.5%. However, the algorithm
           has not been tested with data from polar regions where
           clouds are mainly in the form of ice. Large errors are
           possible for such regions and modifications may be
           This algorithm only requires three input parameters: the
           surface upwelling longwave radiation, the total
           precipitable water vapor and column cloud liquid water.
           The column cloud liquid water used in this study is
           measured by surface microwave radiometer. It would be
           beneficial if cloud liquid water were more accurately
           measured by satellite in the future.
    Title: An analysis of observation and GCM simulations of
           seasonal cycle of monsoon climate in Southeast Asia and
           Tibet Plateau
   Pub No: 9967726
   Author: Zhu, Li
   Degree: PhD
     Date: 2000
    Pages: 185
  Adviser: Bakhru, Hass
     ISBN: 0-599-72446-3
   Source: DAI-B 61/04, p. 1998, Oct 2000
 Abstract: Atmospheric general circulation models have been used for
           climate study. Evaluation of the GCM ability to simulate
           the observed climate features can identify the models'
           inadequacy so that further improvement can be made. The
           research diagnoses GCM simulation of East Asian Climate,
           and is a part of a subproject in the Atmospheric Model
           Intercomparison Program and Coupled Model Intercomparison
           Program. We used three AMIP models and two CMIP models.
           Four land and oceanic regions in East Asia are chosen for
           the diagnostic study. In addition, the Tibetan Plateau is
           included due to its importance to the monsoon system.
           Monthly observational data and GCM simulations for the
           AMIP period 1979&ndash;1993 are used. The work focuses on
           seasonal cycles of surface temperature and precipitation,
           although other climate parameters such as outgoing
           longwave radiation, wind and surface energy balance
           components are also examined.
           ECHAM4 seasonal cycles fit observation the best, with
           correlation coefficients above 0.9 for precipitation and
           almost 1.0 for surface temperature. Other model
           performances are fair. NTU-GCM can simulate particular
           detailed seasonal cycles: the precipitation summer
           retreats in the South China Sea and the Indo-China
           Peninsula. But at the same time its precipitation cycle
           is 78% larger in the Bay of Bengal, showing its
           inconsistency for different domains. NCAR-CCM3
           systematically underestimates the seasonal cycle strength
           in the oceanic domains. Particularly, it underestimates
           the precipitation seasonal cycle 75% in the South China
           Sea and the Tropical Western Pacific.
           NCAR-CSM simulates SST seasonal cycles well compared with
           observation, within 5% discrepancy and improves
           substantially the summer precipitation of CCM3.
           Especially in the South China Sea and the Tropical
           Western Pacific, CSM has summer precipitation of 11.6 and
           12.2 mm/day, compared with observation 11.1 and 9.4
           mm/dcy (CCM3 has 5.37 and 5.49 mm/day respectively). The
           improvement in summer precipitation is mainly due to
           improved surface wind (from 2.6 ms<super>
           &minus;1</super> in CCM3 to 10.4
           ms<super>&minus;1</super> in CSM, which enhances the
           latent heat flux). Therefore the effect of air-sea
           interaction plays an important role in NCAR GCM.
           However, the comparison of MPI GCMs does not show the
           similar improvement for including the air-sea