The research presented here informs the confidence that can be placed in the simulations of land-surface models (LSMs). After introducing a method for simplifying a complex, heterogeneous land-cover dataset for use in LSMs, I show that LSMs can realistically represent the spatial distribution of heterogeneous land-cover processes (e.g., biogenic emission of volatile organic compounds) in Texas. LSM-derived estimates of biogenic emissions are sensitive (varying up to a factor of 3) to land-cover data, which is not well constrained by observations. Simulated emissions are most sensitive to land-cover data in eastern and central Texas, where tropospheric ozone pollution is a concern. I further demonstrate that interannual variation in leaf mass is at least as important to variation in biogenic emissions as is interannual variation in shortwave radiation and temperature. Model estimates show that more-humid regions with less year-to-year variation in precipitation have lower year-to-year variation in biogenic emissions: as modeled mean emissions increase, their mean-normalized standard deviation decreases. I evaluate three parameterizations of subsurface hydrology in LSMs (with (1) a shallow, 10-layer soil; (2) a deeper, many-layered soil; and (3) a lumped aquifer model) under increasing parameter uncertainty. When given their optimal parameter sets, all three versions perform equivalently well when simulating monthly change in terrestrial water storage. The most conceptually realistic model is least sensitive to errant parameter values. However, even when using the most conceptually realistic model, parameter interaction ensures that knowing ranges for individual parameters is insufficient to guarantee realistic simulation. LSMs are often developed and evaluated at data-rich sites but are then applied in regions where data are sparse or unavailable. I present a framework for model evaluation that explicitly acknowledges perennial sources of uncertainty in LSM simulations (e.g., parameter uncertainty, meteorological forcing-data uncertainty, evaluation-data uncertainty) and that evaluates LSMs in a way that is consistent with models’ typical application. The model performance score quantifies the likelihood that a representative ensemble of model performance will bracket observations with high skill and low spread. The robustness score quantifies the sensitivity of model performance to parameter error or data error. The fitness score ranks models’ suitability for broad application. / text

Groundwater is a major source of water in the western US. However, there are limited recharge estimates in this region due to the complexity of recharge processes and the challenge of direct observations. Land surface Models (LSMs) could be a valuable tool for estimating current recharge and projecting changes due to future climate change. In this study, simulations of three LSMs (Noah, Mosaic and VIC) obtained from the North American Land Data Assimilation System (NLDAS-2) are used to estimate potential recharge in the western US. Modeled recharge was compared with published recharge estimates for several aquifers in the region. Annual recharge to precipitation ratios across the study basins varied from 0.01% to 15% for Mosaic, 3.2% to 42% for Noah, and 6.7% to 31.8% for VIC simulations. Mosaic consistently underestimates recharge across all basins. Noah captures recharge reasonably well in wetter basins, but overestimates it in drier basins. VIC slightly overestimates recharge in drier basins and slightly underestimates it for wetter basins. While the average annual recharge values vary among the models, the models were consistent in identifying high and low recharge areas in the region. Models agree in seasonality of recharge occurring dominantly during the spring across the region. Overall, our results highlight that LSMs have the potential to capture the spatial and temporal patterns as well as seasonality of recharge at large scales. Therefore, LSMs (specifically VIC and Noah) can be used as a tool for estimating future recharge in data limited regions.

Land-surface models (LSMs) are crucial components of the Earth system models (ESMs) that are used to make coupled climate-carbon cycle projections for the 21st century. The Joint UK Land Environment Simulator (JULES) is the land-surface model used in the climate and weather forecast models of the UK Met Office. JULES is also extensively used offline as a land-surface impacts tool, forced with climatologies into the future. In this study, JULES is automatically differentiated with respect to JULES parameters using commercial software from FastOpt, resulting in an analytical gradient, or adjoint, of the model. Using this adjoint, the adJULES parameter estimation system has been developed to search for locally optimum parameters by calibrating against observations. This thesis describes adJULES in a data assimilation framework and demonstrates its ability to improve the model-data fit using eddy-covariance measurements of gross primary productivity (GPP) and latent heat (LE) fluxes. The adJULES system is extended to have the ability to calibrate over multiple sites simultaneously. This feature is used to define new optimised parameter values for the five plant functional types (PFTs) in JULES. The optimised PFT-specific parameters improve the performance of JULES at over 85% of the sites used in the study, at both the calibration and evaluation stages. The new improved parameters for JULES are presented along with the associated uncertainties for each parameter. The results of the calibrations are compared to structural changes and used in a cluster analysis in order to challenge the PFT definitions in JULES. This thesis concludes with simple sensitivity studies which assess how the calibration of JULES has affected the sensitivity of the model to CO2-induced climate change.

Remote sensing of land surface temperature (LST) is a complex task. From a satellite-based perspective the radiative properties of the land surface and the atmosphere are inextricably linked. Knowledge of both is required if one is to accurately measure the temperature of the land surface from a space-borne platform. In practice, most satellite-based sensors designed to measure LST over the surface of the Earth are polar orbiting. They scan swaths of the order of 2000 km, utilizing zenith angles of observation of up to 60°. As such, satellite viewing geometry is important when comparing estimates of LST between different overpasses of the same point on the Earth's surface. In the case of the atmosphere, the optical path length through which the surfaceleaving radiance propagates increases with increasing zenith angle of observation. A longer optical path may in turn alter the relative contributions which molecular absorption and emission processes make to the radiance measured at the satellite sensor. A means of estimating the magnitudes of these radiative components in relation to the viewing geometry of the satellite needs to be developed if their impacts on the at-sensor radiance are to be accurately accounted for. The problem of accurately describing radiative transfer between the surface and the satellite sensor is further complicated by the fact that the surface-leaving radiance itself may also vary with sensor viewing geometry. Physical properties of the surface such as emissivity are known to vary as the zenith angle of observation changes. The proportions of sunlit and shaded areas with the field-of-view of the sensor may also change with viewing geometry depending on the type of cover (eg vegetation), further impacting the surface emissivity. / Investigation of the change in surface-leaving radiance as the zenith angle of observation varies is then also important in developing a better understanding of the radiative interaction between the land surface and the atmosphere. The work in this study investigates the atmospheric impacts using surface brightness temperature measurements from the ATSR-2 satellite sensor in combination with atmospheric profile data from radiosondes and estimates of the downwelling sky radiance made by a ground-based radiometer. A line-by-line radiative transfer model is used to model the angular impacts of the atmosphere upon the surfaceleaving radiance. Results from the modelling work show that if the magnitude of the upwelling and downwelling sky radiance and atmospheric transmittance are accurately known then the surface-emitted radiance and hence the LST may be retrieved with negligible error. Guided by the outcomes of the modelling work an atmospheric correction term is derived which accounts for absorption and emission by the atmosphere, and is based on the viewing geometry of the satellite sensor and atmospheric properties characteristic of a semi-arid field site near Alice Springs in the Northern Territory (Central Australia). Ground-based angular measurements of surface brightness temperature made by a scanning, self calibrating radiometer situated at this field site are then used to investigate how the surface-leaving radiance varies over a range of zenith angles comparable to that of the ATSR-2 satellite sensor. / Well defined cycles in the angular dependence of surface brightness temperature were observed on both diumal and seasonal timescales in these data. The observed cycles in surface brightness temperature are explained in terms of the interaction between the downwelling sky radiance and the angular dependence of the surface emissivity. The angular surface brightness temperature and surface emissivity information is then applied to derive an LST estimate of high accuracy (approx. 1 K at night and 1-2 K during the day), suitable for the validation of satellite-derived LST measurements. Finally, the atmospheric and land surface components of this work are combined to describe surface-atmosphere interaction at the field site. Algorithms are derived for the satellite retrieval of LST for the nadir and forward viewing geometries of the ATSR-2 sensor, based upon the cycles in the angular dependence of surface brightness temperature observed in situ and the atmospheric correction term developed from the modelling of radiative transfer in the atmosphere. A qualitative assessment of the performance of these algorithms indicates they may obtain comparable accuracy to existing dual angle algorithms (approx. 1.5 K) in the ideal case and an accuracy of 3-4 K in practice, which is limited by knowledge of atmospheric properties (eg downwelling sky radiance and atmospheric transmittance), and the surface emissivity. There are, however, strong prospects of enhanced performance given better estimates of these physical quantities, and if coefficients within the retrieval algorithms are determined over a wider range of observation zenith angles in the future.

Wind transport and sublimation of snow particles are common phenomena across high altitude and latitude cold regions and play important roles in hydrological and atmospheric water and energy budgets. In spite of this, blowing snow processes have not been incorporated in many mesoscale hydrological models and land surface schemes. A physically based blowing snow model, the Prairie Blowing Snow Model (PBSM), initially developed for prairie environments was used to model snow redistribution and sublimation by wind over two sites representative of mountainous regions in Canada: Fisera Ridge in the Rocky Mountain Front Ranges in Alberta, and Granger Basin in the Yukon Territory. Two models were used to run PBSM: the object-oriented hydrological model, Cold Regions Hydrological Modelling Platform (CRHM) and Environment Canadas hydrological-land surface scheme, Modélisation Environmentale Communautaire Surface and Hydrology (MESH). PBSM was coupled with the snowcover energy and mass-balance model (SNOBAL) within CRHM. Blowing snow algorithms were also incorporated into MESH to create MESH-PBSM. CRHM, MESH and MESH-PBSM were used to simulate the evolution of snowcover in hydrological response units (HRUs) over both Fisera Ridge and Granger Basin.<p> To test the models of blowing snow redistribution and ablation over a relatively simple sequence of mountain topography, simulations were run from north to south over a linear ridge in the Canadian Rocky Mountains. Fisera Ridge snowcover simulations with CRHM were performed over two winters using two sets of wind speed forcing: (1) station observed wind speed, and (2) modelled wind speed from a widely applied empirical, terrain-based windflow model. Best results were obtained when using the site meteorological station wind speed data. The windflow model performed poorly when comparing the magnitude of modelled and observed wind speeds. Blowing snow sublimation, snowmelt and snowpack sublimation quantities were considerably overestimated when using the modelled wind speeds. As a result, end-of-winter snow accumulation was considerably underestimated on windswept HRUs. MESH and MESH-PBSM were also used to simulate snow accumulation and redistribution over these same HRUs. MESH-PBSM adequately simulated snow accumulation in the HRUs up until the spring snowmelt period. MESH without PBSM performed less well and overestimated accumulation on windward slopes and the ridge top whilst underestimating accumulation on lee slopes. Simulations in spring were degraded by a large overestimation of melt by MESH. The early and overestimated melt warrants a detailed examination that is outside the scope of this thesis.<p> To parameterize snow redistribution in a mountain alpine basin, snow redistribution and sublimation by wind were calculated for three winters over Granger Basin using CRHM. Snow transport fluxes were distributed amongst HRUs using inter-HRU snow redistribution allocation factors. Three snow redistribution schemes of varying complexity were evaluated. CRHM model results showed that end-of-winter snow accumulation can be most accurately simulated when the inter-HRU snow redistribution schemes take into account wind direction and speed and HRU aerodynamic characteristics, along with the spatial arrangement of HRUs in the catchment. As snow transport scales approximately with the fourth power of wind speed (u4), inter-HRU snow redistribution allocation factors can be established according to the predominant u4 direction over a simulation period or can change at each time step according to an input measured wind direction. MESH and MESH-PBSM were used to simulate snow accumulation and ablation over these same HRUs. MESH-PBSM provided markedly better results than MESH without blowing snow algorithms.<p> That snow redistribution by wind can be adequately simulated in computationally efficient HRUs over mountainous terrain has important implications for representing snow transport in large-scale hydrology models and land surface schemes. Snow redistribution by wind caused mountain snow accumulation to vary from 10% to 161% of seasonal snowfall within a headwater catchment in the Canadian Rocky Mountains, and blowing snow sublimation losses ranged from 10 to 37% of seasonal snowfall.

Wind transport and sublimation of snow particles are common phenomena across high altitude and latitude cold regions and play important roles in hydrological and atmospheric water and energy budgets. In spite of this, blowing snow processes have not been incorporated in many mesoscale hydrological models and land surface schemes. A physically based blowing snow model, the Prairie Blowing Snow Model (PBSM), initially developed for prairie environments was used to model snow redistribution and sublimation by wind over two sites representative of mountainous regions in Canada: Fisera Ridge in the Rocky Mountain Front Ranges in Alberta, and Granger Basin in the Yukon Territory. Two models were used to run PBSM: the object-oriented hydrological model, Cold Regions Hydrological Modelling Platform (CRHM) and Environment Canadas hydrological-land surface scheme, Modélisation Environmentale Communautaire Surface and Hydrology (MESH). PBSM was coupled with the snowcover energy and mass-balance model (SNOBAL) within CRHM. Blowing snow algorithms were also incorporated into MESH to create MESH-PBSM. CRHM, MESH and MESH-PBSM were used to simulate the evolution of snowcover in hydrological response units (HRUs) over both Fisera Ridge and Granger Basin.<p> To test the models of blowing snow redistribution and ablation over a relatively simple sequence of mountain topography, simulations were run from north to south over a linear ridge in the Canadian Rocky Mountains. Fisera Ridge snowcover simulations with CRHM were performed over two winters using two sets of wind speed forcing: (1) station observed wind speed, and (2) modelled wind speed from a widely applied empirical, terrain-based windflow model. Best results were obtained when using the site meteorological station wind speed data. The windflow model performed poorly when comparing the magnitude of modelled and observed wind speeds. Blowing snow sublimation, snowmelt and snowpack sublimation quantities were considerably overestimated when using the modelled wind speeds. As a result, end-of-winter snow accumulation was considerably underestimated on windswept HRUs. MESH and MESH-PBSM were also used to simulate snow accumulation and redistribution over these same HRUs. MESH-PBSM adequately simulated snow accumulation in the HRUs up until the spring snowmelt period. MESH without PBSM performed less well and overestimated accumulation on windward slopes and the ridge top whilst underestimating accumulation on lee slopes. Simulations in spring were degraded by a large overestimation of melt by MESH. The early and overestimated melt warrants a detailed examination that is outside the scope of this thesis.<p> To parameterize snow redistribution in a mountain alpine basin, snow redistribution and sublimation by wind were calculated for three winters over Granger Basin using CRHM. Snow transport fluxes were distributed amongst HRUs using inter-HRU snow redistribution allocation factors. Three snow redistribution schemes of varying complexity were evaluated. CRHM model results showed that end-of-winter snow accumulation can be most accurately simulated when the inter-HRU snow redistribution schemes take into account wind direction and speed and HRU aerodynamic characteristics, along with the spatial arrangement of HRUs in the catchment. As snow transport scales approximately with the fourth power of wind speed (u4), inter-HRU snow redistribution allocation factors can be established according to the predominant u4 direction over a simulation period or can change at each time step according to an input measured wind direction. MESH and MESH-PBSM were used to simulate snow accumulation and ablation over these same HRUs. MESH-PBSM provided markedly better results than MESH without blowing snow algorithms.<p> That snow redistribution by wind can be adequately simulated in computationally efficient HRUs over mountainous terrain has important implications for representing snow transport in large-scale hydrology models and land surface schemes. Snow redistribution by wind caused mountain snow accumulation to vary from 10% to 161% of seasonal snowfall within a headwater catchment in the Canadian Rocky Mountains, and blowing snow sublimation losses ranged from 10 to 37% of seasonal snowfall.

Land surface covers only 30% of the global surface, but contributes largely to the intricacy of the climate system by exchanging water and energy with the overlying atmosphere. The partitioning of incident solar radiation among various components at the land surface, especially vegetation and underlying soil, determines the energy absorbed by vegetation, evapotranspiration, partitioning between surface sensible and latent heat fluxes, and the energy and water exchange between the land surface and the atmosphere. Because of its significance in climate model, land surface model solar radiation partitioning scheme should be evaluated in order to ensure its accuracy in reproducing these naturally complicated processes. However, few studies evaluated this part of climate model. This study examines a land surface solar radiation partitioning scheme, i.e., that of the Community Land Model version 4 (CLM4) with coupled carbon and nitrogen cycles. Taking advantage of multiple remote sensing fraction of absorbed photosynthetically active radiation (FPAR) datasets, ground observations and a unique 28-year FPAR dataset derived from the Global Inventory Modeling and Mapping Studies (GIMMS) Normalized Difference Vegetation Index (NDVI) dataset, we evaluated the CLM4 FPAR’s seasonal cycle, diurnal cycle, long-term trends and spatial patterns. Our findings show the model roughly agrees with observations in the seasonal cycle , long-tern trend and spatial patterns but does not reproduce the diurnal cycle. Discrepancies also exist in seasonality magnitudes, peak value months and spatial heterogeneity. We identified the discrepancy in the diurnal cycle as due to the absence of dependence on sun angle in the model. Implementation of sun angle dependence in a one-dimensional (1-D) model is proposed. The need for better relating vegetation to climate in the model indicated by long-term trends is also noted. Evaluation of the CLM4 land surface solar radiation partitioning scheme using remote sensing and site level FPAR datasets provides targets for future development in its representation of this naturally complicated process. / text

Land Surface Parameterization (LSP) schemes have evolved from simple tipping-bucket models to fully interactive models, including parameterizations which account for exchanges of momentum, energy, mass, and biogeochemistry. As the demand for greater realism has increased, so has the complexity of LSPs which now includes some parameters that may not be universally relevant to all regions of the globe. The performance of LSP schemes depends on the magnitude of structural, data-related (input and output), and parameter uncertainties in the model. Parameter estimation uncertainty can be reduced by calibrating LSPs against measurements available at field sites. Given the multiple outputs of the models, multi-objective optimization approaches are performed. Some of the parameter values used in LSPs have originally obtained from laboratory studies which analyzed plant behavior under a range of conditions in enclosed chambers. The research described in this dissertation takes advantage of currently available data from several eddy covariance flux towers located mainly in the Brazilian Amazon basin to estimate parameter values of a widely-used LSP scheme, version 3 of the Simple Biosphere model (SiB3). Background climatological data was used to assess the representativeness of the data collection period that might have affected model calibration. Variance-based sensitivity analysis was then used to investigate potential structural deficiencies in SiB3 and to reduce the dimensionality of the subsequent optimization by identifying those model parameters that merit calibration. Finally, some structural and conceptual aspects of SiB3 were tested inside Biosphere 2 Tropical Rain Forest biome (B2-TRF) under meteorological conditions that resemble those predicted in future climate scenarios for the Amazon basin.

主催：JST/CREST，Vrije University, ALTERRA, IBPC

This study implemented a GIS-based analysis of local climate zones (LCZ) in Denton, TX with data sets from 2009, 2011, 2015, and 2016. The LCZ scheme enables evaluation of distinct regions' thermal characteristics with greater granularity than conventional urban-rural dichotomies. Further, the GIS-based approach to LCZ mapping allows use of high-resolution lidar data, the availability of which for the study area enabled estimation of geometric and surface cover parameters: height of roughness elements, sky-view factor, and building surface fraction. Pervious surface fraction was estimated from National Landcover Database impervious imagery. A regular grid was used to estimate per-cell mean values for each parameter, and with a decision-making algorithm (if/then statements) two maps were produced (2011 and 2015) and six LCZ identified in each: LCZ 6 (open low-rise), LCZ 8 (large low-rise), LCZ 9 (sparsely built), LCZ A (dense trees), LCZ B (scattered trees), and LCZ C (bush/scrub). Post-processing was carried out to ensure identified zones met the spatial minimum for qualification as LCZ. Landsat Collection 2 Level 2 surface temperature products from various seasons of 2011 and 2015 were acquired to examine LCZ thermal differentiability, and preliminary surface urban heat island intensity values were estimated. Particular attention was afforded to issues regarding data quality and classifier threshold adjustment.