H13C-0415 1340h
Distributed Model for Overland Hydrology, Sediment and Chemical Transport in Watersheds
Contaminants from non-point or distributed sources can adversely affect the quality of downstream waters and groundwater. Reliable models can help understanding the fate and transport of contaminants on the overland as well as the shallow subsurface and lead to better management strategies of the environmental response and security. In this study, a watershed is represented by discretizing contributing areas into a cascade of one dimensional (1D) overland flow using DEM data. A distributed watershed model is developed to simulate overland flow and processes of sediment and chemical detachment and their transport in runoff. The model, consisting of linked hydrology, sediment and chemical transport components, is physically based and process oriented. In the model, the overland flow component uses fully implicit nonlinear finite difference scheme combined with a linear scheme to numerically solve the 1D kinematic wave equations. The erosion and sediment transport component simulates the movement of eroded soil along with the movement of surface water, calculates the sediment transport capacity and the potential sediment supply, chooses whichever is less and routes it throughout the watershed. Erosion is the sum of splash detachment by raindrops and flow detachment by runoff. The process of chemical transport in the soluble phase is simulated by a 1D advection, diffusion and adsorption-desorption processes equation. Chemical transport by suspended sediment is described by the advection-diffusion equations with the sink-source term including erosion-deposition exchange processes. Both sediment and chemical transport components are developed as separate blocks of the watershed overland modeling system. A 1D model in overland studies can provide plenty of useful information without requiring many input parameters. The model will be incorporated into the Army Risk Assessment Modeling System (ARAMS) as a watershed module to satisfy the need for simulating hydrological, sediment, and chemical transport at different temporal and spatial scales.
H13C-0416 1340h
Effects of Implementing MODIS Land Cover Data in MM5
A new land-cover classification from the Moderate-resolution Imaging-Spectroradiometer (MODIS) was incorporated in the fifth-generation Pennsylvania State University-National Center for Atmospheric Research (PSU-NCAR) Mesoscale Model (MM5). Surface parameter sets are determined by translating 17-category MODIS classes into the U.S. Geological Survey (USGS) and Simple Biosphere (SiB) categories available in MM5. MODIS surface albedo was also used in this study as one of the surface parameters. Current MM5 relies on 1-km global land-cover map produced using data from the National Oceanic and Atmospheric Administration's Advanced Very High Resolution Radiometer (AVHRR) from 1992 to1993. By updating this last decade changes in land cover from MODIS, their impacts on the regional diagnosis of near-surface atmospheric state variables as well as characteristics of the planetary boundary layer were investigated. Simulation results show that MODIS land-cover replacement showed significant influence on the magnitudes of air temperature (up to ±5°C) and humidity (up to ±1gkg-1). Area-averaged comparisons with ground measurements showed a small but persistent improvement with a further enhancement on a clear-sky day when MM5 was simulated with MODIS land-cover classification. Such improvements were more significant at a localized scale. MODIS albedo addition did not alter the results much obtained with the inclusion of the MODIS land-cover map. This suggests that the model needs more spin up time to adjust new albedo values.
H13C-0417 1340h
Validation of Satellite-Based Nesdis Rainfall Products
Evaluating the quality of any satellite-based rainfall product is useful and required for improving its algorithm. The objective of this study is to develop a statistical approach for evaluating the satellite-based NESDIS rainfall products from Hydro-Estimator (HE) algorithm, GOES Multi-Spectral Rainfall Algorithm with Daytime and Nighttime Rain Screen (GMSRA\#2), and IR/microwave Blended Algorithm (Blend). Capability of each NESDIS rainfall product is examined with respect to seasonal variability, climate conditions, and storm types over different topographic regions. Ground-based radar and gauge rainfall observations, at high resolution (hourly), are used for validating NESDIS product. In this study, high resolution NESDIS product (hourly 4 Km $\times$ 4 Km) is evaluated regionally in details over several small size study sites with high hourly (daily when hourly is not available) rain-gauge density, for instance: specific $1\deg$ $\times$ $1\deg$ degrees. The size of every study site varies from $0.5\deg$ $\times$ $0.5\deg$ to $2\deg$ $\times$ $2\deg$ degrees, depending on the density and distribution of available hourly rain-gauge stations over the study site and storm size. Study time period is from April 2003 to present, when archived NESDIS rainfall products are available. Three NESDIS rainfall products have been compared with NEXRAD Stage-IV rainfall, rain gauge observations, and GOES infrared (IR) cloud images for four storms during the winter and summer seasons over a $2\deg$ $\times$ $2\deg$ validation study area located in Hernando County, Florida ($28\deg$N-$30\deg$N and $81\deg$W-$83\deg$W). The first test (I) was conducted for a six-hour storm event that took place on 02/24/2004 from 14:00 to 20:00 UTC, the second test (II) was for a six-hour storm that took place on 03/16/2004, from 00:00 to 06:00 UTC, the third test (III) was for a six-hour storm that took place on 08/22/2003, from 18:00 to 24:00 UTC, and the forth test (IV) was for a six-hour storm that took place on 09/03/2003, from 18:00 to 24:00 UTC. The primary results from tests I and II demonstrated that HE estimates rainfall with lower intensity and also over different locations compared to NEXRAD rainfall during the winter but, from tests III and IV, a remarkable improvement has been noticed in the performance of the algorithm which was also confirmed when compared to the rain gauge observations and the GOES infrared cloud images. Those preliminary results confirmed the fact that Hydro-Estimator (HE) has shown a substantially improved ability to estimate precipitation compared to GMSRA\#2 and Blend algorithms.
H13C-0418 1340h
Fire Forcing of Catchment Erosion in Burned Areas of Saint Gabriel Mountains, Southern California
A methodology to evaluate fire flood risk, connected to catchment erosion, in wildfires and high rainfall prone areas, by coupling deterministic and statistic analysis, is proposed. Accordingly, soil physical-chemical fire-induced changing, altering hydrological and sedimentological response, are kept in account via hydrological-sedimentological distributed model. Also, probability for assigned frequency of peak flows and sediment yields is estimated for different scenarios (i.e. burned and unburned). Application to 9 San Gabriel Mountains basins, shows this methodology able to describe a wide range of field measurements, collected in natural (pre-fire) and altered (post-fire) conditions. The method is also useful to give an explanation in the probability domain of the hydrological and sedimentological enhanced risk due to the presence of wildfires.
H13C-0419 1340h
A Quantitative and Objective Procedure for Evaluation of Distributed Hydrologic and Hydrometeorological Models
Better spatial pattern representation of hydrological variables is one of the important targets of distributed hydrologic and hydrometeorological models at basin scale. In this paper, we introduce a set theory-based similarity measure - the Hausdorff Norm (HN), for a quantitative evaluation of distributed fields of hydro-meteorological variables against observations. The procedure can be used for the evaluation of individual fields or for the evaluation of several fields simultaneously. The procedure is not limited to snapshots as a sequence of several distributed fields in time can also be evaluated. The procedure is tested using the noah LSM over a semi-arid watershed, the San Pedro River basin in Arizona. The noah model is run using three different uniform resolution grids (1, 4, 12 km) and a multiple resolution grid scheme (1 km to 12 km) defined by the degree of land surface and sub-surface characteristics of this watershed. The model is driven in offline fashion with hourly forcing data remapped from the NLDAS. Surface characteristics like vegetation, soil and topography are remapped from a fine resolution uniform grid (1km). An evaluation of the influence of vegetation parameter estimation is also carried out. Vegetation parameters were estimated at multiple representative sites in the San Pedro basin using a multi-objective framework. All the evaluations are carried out for July 2003 (a period of strong monsoon influence). The ground surface temperature fields are contrasted to those obtained from the MODIS-AQUA instrument. For the simultaneous multiple field evaluation/comparison we used the results from the NLDAS Mosaic model. For the simultaneous multiple field evaluation/comparison of the different distributed noah model resolutions and the evaluation of the influence of the parameter estimation procedures a 1 km noah model resolution output using the non-optimized parameter values (model default) is the benchmark. These extensive and comprehensive tests have shown the extreme versatility and power of the HN as a tool for evaluation and comparison of distributed model performances at different spatial scales. The HN provides with a quantitative measure of the model goodness.
H13C-0420 1340h
Integrating Remotely Sensed Estimates of Evapotranspiration Into Long-Term Ground Measurements, Three Gorges Region of China
Accurate estimates of local and regional evapotranspiration are essential for determining water and energy budgets. Data are rare for many under-gauged regions of the world, so the use of remote sensing data to determine these estimates is especially appealing. Remote sensing offers greater spatial coverage (i.e. global) than traditional measures, but has its limitations (e.g., may have less-frequent temporal sampling, and the length of such datasets is limited to the launch of appropriate satellites) and uncertainties associated with the derived evapotranspiration estimates. In this study, we apply the method by Liang and Islam (2001) to estimate the evapotranspiration (ET) for a large area using remote sensing information (e.g., MODIS data). ET estimates from this approach will be compared to the evapotranspiration data measured by two methods (small and large pans) on the ground. Evaluations of the validity of the method by Liang and Islam and other relevant methods for the study region will be provided. The area to be studied is in the Three Gorges Region of China, a region undergoing rapid environmental, land-use, and hydrological change. As the Three Gorges Dam undergoes filling, accurate assessments of the region's hydrology have important implications for its water quality, public health, and economy.
H13C-0421 1340h
Implementing a new multi-scale flow network routing scheme in a land surface model
This paper first presents a new multi-scale approach to generate the flow network for land surface models at different spatial scales. The new approach has two advantages: (1) it allows runoff in a land surface model grid to exit through multiple directions simultaneously rather than through only one of the eight discrete directions as in many other methods (i.e., either to adjacent or diagonal neighbor according to the steepest downward slope); and (2) it introduces a concept of the elastic coefficient to determine hydrologic parameters for more accurate flow routing across different spatial scales. The flow network generated by the new multi-scale approach in conjunction with a kinematic wave model is fully coupled and implemented in the VIC-3L (Three-Layer Variable Infiltration Capacity) model to test the applicability of the new approach for land surface models across scales. The watershed of Illinois River at Gore in Oklahoma is selected for the case study. Improvements on streamflow simulations by the new multi-scale approach and the impacts on evapotranspiration and soil moisture simulations will be presented and discussed at different spatial resolutions. The new multi-scale flow network generation approach can be easily implemented in other land surface models.
H13C-0422 1340h
The Passive Surface Water Fluxmeter to Measure Cumulative Water and Solute Mass Fluxes
Flux and discharge based data are increasingly being viewed as critical information needed to address various components of contaminant hydrology such as source prioritization, risk prediction, compliance monitoring, remediation endpoint evaluation, and contaminant attenuation assessment. Numerous methods exist to measure instantaneous or cumulative local water flow velocities (fluxes) and global discharges in stream transects. Measurements of solute mass discharges are currently performed by measuring solute concentrations at discrete moments in time and discrete locations across a stream transect. These measurements have to be combined with the independently obtained water flow velocities or discharges to arrive at estimates of local instantaneous solute mass fluxes. In general, water flux and solute concentration data are not obtained at coincident points in time or space, thus introducing some conceptual error. Furthermore, cumulative solute mass discharges are only obtained from a temporal and spatial interpolation and integration of the data. The method proposed here makes use of a Passive Surface Water Fluxmeter (PSFM), which is capable of directly measuring local cumulative or time averaged water and solute mass fluxes in surface water flow systems. This is, the PSFM gives direct estimates of solute mass fluxes without requiring independent water flux and concentration measurements. Furthermore, due to its passive nature, it inherently performs the time integration of local water and solute mass fluxes. Several PSFMs can be deployed along a stream transect, where each PSFM delivers data at a number of depths. This results in a point matrix of cumulative flux measurements that can be interpolated and integrated to arrive at estimates of the global water and solute mass discharges. Results of first laboratory flume experiments with the PSFM will be presented.
H13C-0423 1340h
Use of a Spatially-Distributed, Process-Based Hydrologic Model to Simulate the Influence of Discontinuous Permafrost on Hydrologic Processes
In the sub-arctic environment, permafrost is a strong factor in controlling many hydrologic processes including stream flow and soil moisture. Soil moisture, which displays a high spatial and temporal variability, is an important variable in understanding and predicting a large number of processes including land-atmosphere interactions and permafrost aggradation/degradation. In order to understand and predict ecosystem response to a changing climate and resulting feedbacks, it is critical to quantify the interaction of soil moisture and meteorology as a function of climatic processes, landscape type, and vegetation. The primary goal of our research is to describe, simulate, and predict soil moisture dynamics and all other hydrologic processes everywhere throughout a sub-arctic watershed. The model we are developing, TopoFlow, is being used as a tool to better understand the effects of vegetation and soil type, presence or absence of permafrost, the amount and timing of precipitation, and disturbance (such as wildfire) on soil moisture dynamics. Three small sub-basins of the Caribou-Poker Creeks Research Watershed (CPCRW), located 48 km north of Fairbanks, Alaska (65\deg 10'N, 147\deg 30'W), are the areas selected for study. These small sub-basins, which are underlain with approximately 3, 19, and 53% permafrost, are simulated to explore differences in permafrost versus non-permafrost areas. Discontinuous permafrost is represented though differences in hydraulic conductivity between the permafrost and non-permafrost soils. Non-permafrost soils, or soils within the seasonally thawed soils, are represented with much larger hydraulic conductivities than in permafrost soils. The primary control on local hydrological processes is dictated by the presence or absence of permafrost, but is also influenced by the thickness of the active layer and the total thickness of the underlying permafrost. As permafrost becomes thinner or decreases in areal extent, the interaction of surface and sub-permafrost ground water processes becomes more important. The inability of soil moisture to infiltrate to deeper groundwater zones due to ice rich permafrost maintains very wet soils in arctic regions. However, in the slightly warmer regions of the sub-arctic, the permafrost is thinner or discontinuous. In permafrost-free areas, surface soils can be quite dry as infiltration is not restricted, impacting ecosystem dynamics, fire frequency and latent and sensible heat fluxes.
H13C-0424 1340h
A Simple method for Spatial Disaggregation of Radiometer Derived Soil Moisture Using Higher Resolution Radar Observations
This paper presents a technique for estimation of soil moisture by combining radiometric brightness temperatures in the LH band with horizontally co polarized L band radar backscattering coefficients. The approach is to use radiometric estimates of soil moisture at a lower resolution to compute the sensitivity of radar to soil moisture at the lower resolution. This estimate of sensitivity is then disaggregated using vegetation water content, vegetation type and soil texture information, which are the parameters on which radar sensitivity to soil moisture depends and are generally available at a scale of radar observation. The method discussed in this paper has potential applicability in soil moisture retrieval from proposed passive/active L band satellite instruments. The HYDROS instrument is proposed to have an L band radiometer and L band radar onboard. The passive instrument will have spatial resolution of the order of tens of kilometers and will operate along with the active instrument that will take observations at a resolution of tens of meters. The present study applies the methods presented to a limited data set obtained from the SMEX02 campaign held in June - July 02 in Iowa during which an airborne L band radiometer (PALS) and an L band synthetic aperture radar (AIRSAR) were used to coincidentally acquire data over the same region on 3 days and 400 m and 30 m resolutions respectively. In situ sampling of soil and vegetation parameters was also done. To demonstrate the applicability and limitation of the technique over a period of weeks, a simulated experiment was performed, the results of which have been presented and discussed in this study.
H13C-0425 1340h
Validation and Error Characterization of GPCP-1DD Precipitation Product over the Contiguous United States
A validation and error characterization study of the GPCP-1DD precipitation product over the contiguous United States is presented. Daily precipitation estimates over a 1-degree grid are compared against aggregated precipitation values obtained from the forcing field of the North American Land Data Assimilation System (LDAS). LDAS daily values are consistent with the CPC gauge-based daily precipitation product, and hence can be regarded as realistic ground-truth values with full coverage of the United States. Continuous and categorical measures of skill are presented, so that both the ability of GPCP-1DD to identify a precipitation event and its accuracy in determining cumulative precipitation amounts are evaluated. Daily values are aggregated into seasonal averages, and spatial averages are computed for five arbitrarily defined climate zones that cover most of the study area. Results show that in general there is good agreement between GPCP-1DD and LDAS values, except for particular areas where GPCP-1DD is unable to identify high-intensity events. For those events correctly identified by GPCP-1DD, computation of continuous statistics show that average bias is negligible in most areas of the U.S. except for humid regions north of parallel 40°N. However, the RMSE statistics shows that differences in estimated precipitation for individual 1-degree cells can be significant, exceeding in most cases the magnitude of the average precipitation. Beyond the validation, the error characterization presented here can significantly enhance the utility of the GPCP 1DD product for use in ensemble hydrologic modeling and forecasting.
H13C-0426 1340h
Relation between satellite-derived vegetation indices, surface temperature and vegetation water content
Vegetation and its relation to other climatic variables are very important to understand complex land-atmosphere interactions, which significantly influence climate. In this paper we study the relationship 1) between the surface temperature(Ts) and the normalized difference vegetation index(NDVI), and 2) between NDVI and vegetation water content(VWC) from the remotely sensed data. The monthly TIROS (Television Infrared Observation Satellite) Operational Vertical Sounder(TOVS) dataset and NOAA Advanced Very High Resolution Radiometer(AVHRR) dataset are examined for the relationship between Ts and NDVI respectively for the year 1985 - 1992. This relationship is described in the TvX plots, whose geometry explains the interrelationship between Ts, vegetation cover, and moisture availability. The study areas for this research are the North American Monsoon System (NAMS) (latitude : -120W to -85W, longitude : 10N to 50N) and the international H2O project(IHOP) area (latitude : -115W to -95W, longitude : 32N to 41N). The relationship between vegetation cover and moisture availability evaluates this TvX relationship. In the second part of this research, the daily Moderate-resolution Imaging Spectroradiometer (MODIS) data set and the 2-day Advanced Microwave Scanning Radiometer - Earth Observing System (AMSR-E) data set from March to September 2004 are studied for NDVI and VWC respectively at continental scale. Comparison of these two data sets shows that NDVI and VWC are generally related to each other in most continental scale watersheds.
H13C-0427 1340h
Seasonal Soil Frost in the upper Mississippi River Basin
Seasonal soil frost can play a significant role in controlling energy and water fluxes through the winter and spring in the upper Mississippi River basin. Ice in the soil can impede the infiltration of spring melt and precipitation events, potentially leading to higher peak flows on regional streams. It also reduces soil moisture drainage, minimizing baseflow generation and groundwater recharge through the winter and yielding wetter soils in the spring. In terms of the land-surface energy balance, changes in soil moisture freeze/thaw state requires a significant allocation of energy, which results in a suppression of the magnitude of diurnal surface temperature changes. This leads to a change in the significance of sensible heat fluxes to the surface energy balance. The Variable Infiltration Capacity (VIC) macroscale hydrology model has been used to simulate the effects of seasonal soil frost on water and energy fluxes in the upper Mississippi River basin. Simulations were conducted using daily meteorological observations from 1950 through 1997. Comparisons of water and energy fluxes between model simulations with and without the frozen soil algorithm are used to evaluate the role of seasonal soil frost in the hydrologic cycle and identify seasonal conditions under which the response to frozen soil is more pronounced.
H13C-0428 1340h
HOURLY AND DAYTIME EVAPOTRANSPIRATION FROM GRASSLAND USING RADIOMETRIC SURFACE TEMPERATURES
Estimates of evapotranspiration (ET) are needed for many applications in agriculture, hydrology and meteorology because ground-based measurement techniques of ET and variables controlling it, such as canopy density (i.e., leaf area index LAI), soil water availability, and surface temperature (Ts) are inadequate over large or heterogeneous areas. Remote sensing can be a handy source for such variables at a reasonable resolution. Soil moisture availability is a key variable, as it exerts control over the ratio between actual and potential ET. Although soil moisture sensing is progressing rapidly, remotely sensed soil moisture content data are not always available or accurate, especially for dense vegetation. Moreover, remotely sensed soil moisture does not represent the entire soil water profile (root zone) that controls ET. Therefore, a method is needed to find ET directly from Tsr, without requiring soil water availability. In this study, we propose a procedure to estimate ET using Tsr. The method uses a dimensionless temperature DT, defined as (Tsa - Ta)/(Tmax - Ta), where Tsa is aerodynamic surface temperature, Ta is the air temperature and Tmax is the surface temperature that would occur if all the net radiation (Rn - G) was converted to sensible heat flux (H) and no evaporation occurred. The aerodynamic surface temperature is the temperature that gives the correct value of H at a clearly specified value of the scalar roughness length, zoh, based on Monin-Obukhov Similarity (MOS) theory in the surface sublayer. Radiometric surface temperature is converted into aerodynamic surface temperature using an Analytical Land-Atmosphere Radiometer Model (ALARM). Instantaneous (or hourly) ET was extrapolated to daily ET by assuming a constant evaporative fraction (EF = ET/Rn). This approach has been tested on data taken at two grassland sites. The results demonstrate that, for grassland, the model gives good estimates of ET when Ta and Tsr are available. The method presented has the conceptual advantage that it produces EF, which is a key variable to characterize the hydrology of a site, directly in terms of surface and air temperatures.
H13C-0429 1340h
Application of Unscented Kalman Filter for Higher Order Accuracy in the Assimilation of Near Surface Soil Moisture
Coupled land-atmosphere models are increasingly focused on using assimilated near-surface soil-moisture to improve the prediction of moisture and heat fluxes. Any assimilation scheme requires efficient handling of uncertainties, manifested through observational errors and errors in the land surface model (accounting for inaccuracies in specification of initial conditions, surface boundary parameters, forcing data, underlying model physics etc.), and better representation of inherent non-linearity in the system. The extended Kalman filter (EKF) is one of the well known algorithms for data assimilation in nonlinear systems. Unfortunately, EKF approximates the state variable as a Gaussian random variable (GRV), which then is propagated through the first-order linearization of the nonlinear system. This can seriously affect the accuracy or even lead to the divergence of the nonlinear system, often requiring resetting the state covariance to its positive definiteness. Ensemble Kalman filter (EnKF), which is easier to implement in complex models, provides a sub-optimal alternative. EnKF propagates the probability distribution of the state variable using a randomly selected set of realizations (ensemble) through the nonlinear model. The necessary statistics are then obtained from the ensemble analysis. But the dependence of its performance on random sampling (some statistical features may be lost) and requiring large number of ensemble members for non-Gaussian distributions can make EnKF computationally expensive and sometimes infeasible. The unscented Kalman filter (UKF) addresses this problem through an unscented transformation, wherein the state variable is approximated as a random variable, but it is now represented using a minimal set of sample points. These minimal sample points are deterministically chosen, rather than random sampling (as in Mote-Carlo techniques, EnKF) and completely capture the mean and covariance accurately to at least the 3rd order for any nonlinearity and the higher moments to at least 2nd order, with the same computational efficiency as EKF. The present study evaluates the performance of UKF for the prediction of soil moisture and associated fluxes, and their prediction errors, through implementation in NCAR's Land Surface Model.
H13C-0430 1340h
Feasibility of snow water equivalent estimation using the Ensemble Kalman Filter
Recent studies have demonstrated the tremendous affect of snow cover on the dynamics of land surface behavior. Due to the large latent heat of vaporziation of water and the high albedo of snow, snow cover plays a key role in the global energy balance. Additionally, it is well-known that many arid regions depend upon the seasonal snowpack as a main source of water. These facts have driven the study of snow and the interest in increasingly sophisticated snow characterization within land surface models (LSMs). The large uncertainty associated with the LSMs and with precipitation data in general is one reason why work is being done to constrain the models with observations of snow states and remote sensing observations. The latter have a tremendous potential for improving model performance due to the relative ease of collection when compared with in-situ observations. Attempts have been made to obtain snow water equivalent (SWE) estimates by inverting remote sensing observations. Such methods include empirical regression algorithms based upon results from radiative transfer models (RTMs), `direct insertion' methods for incorporating retrieved snow state estimates directly into an LSM, and neural network-based schemes which adjust snow states until predictions from an RTM match the remote sensing observation. While each of these methods have had some success, the robust estimation of SWE and other snow properties is made difficult by the complex dependence of the remotely sensed signal on a number of snowpack characteristics. Data assimilation is an ideal framework for merging multi-frequency remote sensing observations and a snow physics model, because it provides a means of weighing the tremendous uncertainty of meteorological data (such as precipitation measurements from snow gages) and remote sensing observations. The objective of this study is to assess the feasibility of using the Ensemble Kalman Filter (EnKF) to characterize snowpack properties. The EnKF follows a Monte Carlo simulation approach in which all parameters, forcing variables, and initial conditions are treated as random variables. By simulating the evolution of an ensemble of state variables, an estimate of the uncertainty of the state variables is obtained. The relative uncertainty of the state variables is weighed against the uncertainty in the remote sensing observations through the EnKF update equation. The advantage of the EnKF data assimilation framework is that it easily allows for the incorporation of observations from several different portions of electromagnetic spectrum; each observation may contain different information about the snowpack, is available at a different spatial and temporal resolution, and has a different level of uncertainty. In this study, a season-long, one-dimensional experiment is performed in which synthetic observations from the microwave, visible/near-infrared, and thermal infrared spectra are used to update snowpack states in a land surface model. Specifically, synthetic microwave observations from AMSR-E and SSM/I, as well as synthetic broadband albedo observations, and surface temperature observations from MODIS are assimilated. The land surface model has been modified to compute estimates of the average grain diameter and the snowpack albedo. Various parameters which affect the evolution of the snowpack are treated as random variables in order to make the LSM compatible with ensemble forecasting. Results from the assimilation are compared to those from a pure modeling approach and from a remote sensing inversion approach. Tradeoffs between ensemble size, estimation error, and computational expense are explored.
H13C-0431 1340h
Comparison of Surface Turbulent Flux Estimation from Radiometric Surface Temperature Observations Using Retrieval- and Data Assimilation-Based Approaches
Surface moisture and energy fluxes are extremely important in land-atmosphere interaction. Despite the importance of these fluxes, in-situ measurements are generally limited to sparse ground-based monitoring sites that are often associated with short-term field experiments or to extremely localized regions. To create operational frameworks for estimating these fluxes over longer periods and larger spatial extents will require the use of remote sensing products. While there are many potentially useful remote sensing observations that can be used to estimate land surface fluxes, radiometric surface temperature observations are often used because i) it is a variable that is implicit in the surface energy balance and ii) is measured operationally from several orbiting platforms that provide excellent spatial and temporal resolution and coverage. Techniques used to estimate surface fluxes from radiometric surface temperature generally fall into two categories: retrieval-based and data assimilation approaches. Retrieval methods provide an instantaneous estimate of the variable of interest (e.g. surface fluxes) from the observed quantity (surface temperature) based on the inversion of a physical or empirical model. Data assimilation approaches differ from retrieval methods in that they combine a physically-based model with a sequence of remote sensing observations and attempts to merge the two to produce an optimal estimate based on the relative uncertainty of all inputs. Up to this point, there has been little, if any direct comparison between retrieval- and assimilation-based techniques to assess the side-by-side performance of these two different approaches. In this study we compare the popular triangle retrieval method to a variational data assimilation approach for estimating surface turbulent fluxes from radiometric surface temperature using data from a field site in Kansas. The two methods were applied in parallel for two full summer seasons. Synthetic tests were first performed to assess the ability of the methods to cope with known input errors. Given minimal input errors, the two methods performed comparably, with the data assimilation method appearing more robust as input errors were increased. Additionally, comparison of computational burden showed significant savings by the assimilation approach. The two methods were then applied using real radiometric surface temperature inputs and estimates were compared to ground-based surface flux observations. Results show that both methods are capable of providing reasonable estimates of latent and sensible heat fluxes, with better results from the variational assimilation approach. Additionally, based on the necessary model inputs and assumptions and computational burden, it would appear that the assimilation-based approach has advantages for operational application.
H13C-0432 1340h
Modeling storm runoff reduction through urban forestry
Effects of urban forests on runoff reduction have been conceptualized in management practice, but not well quantified due to lack of practical application tools. In this study, a storm runoff reduction model was developed and applied in an urban watershed in Bismarck, North Dakota. Methods and analyses conducted in this model were based on research embodied in the Small Watersheds Model TR55 (USDA SCS, 1986) and the Single Tree Rainfall Interception Model (Xiao et al, 1998, 2000). This model simulates storm runoff reduction at the individual tree level for different size storm events. It fully considers the effects of microclimate, soil, land cover, tree species, and tree size on runoff reduction. The total amount of storm runoff and amount of runoff reduction is reported. Thus, the model is a management tool for professionals needing to include trees as BMPs in stormwater management programs.
http://cufr.ucdavis.edu
H13C-0433 1340h
All sky Predictions of Solar Irradiance, air Temperature and Humidity Using Agua and Terra MODIS and GOES-10 Satellite Data
With the new generation of earth observation satellites, opportunities for improved spatial and temporal descriptions of key environmental controls on the exchanges of carbon dioxide and energy have emerged. In this study, a satellite based scheme for the retrieval of spatially distributed input of air temperature, air humidity and solar irradiance for all sky conditions is presented using data acquired from the Agua and Terra MODerate resolution Imaging Spectroradiometer (MODIS). MODIS retrievals of atmospheric state and cloud variables are assimilated in a radiative transfer model that solves for direct-beam and diffuse solar irradiance for all sky conditions at an hourly time scale. Cloud-drift vectors from GOES-10 are implemented in order to resolve the diurnal variation in the irradiance components. Clear-sky air and dew-point temperatures are retrieved from MODIS atmospheric profiles and the predictions are extended to all sky conditions and a diurnal time scale from the use of multiple daily MODIS overpasses. The predictions are evaluated against standard meteorological data from a semi-arid grassland site in southeast Arizona for the year 2003, and the encouraging results emphasize the applicability of satellite estimates of key meteorological variables in ecosystem models for a more accurate evaluation of regional or global budgets of carbon dioxide and energy exchange.
H13C-0434 1340h
Estimation of the Spatial Distribution of Surface Soil Moisture Using Non-Stationary Geostatistical Methods
The spatial distribution of surface moisture content within the upper 6-cm soil layer is estimated using both ground-based measurements and non-stationary geostatistical methods within a 50-km by 100-km region in Iowa during the Soil Moisture Experiments in 2002 (SMEX02). The non-stationary portion of the distribution (external drift) is calculated by applying hourly precipitation data in the simplified soil moisture dynamics equation. The covariance function of the residual random portion is derived from the spatial correlations of soil texture and vegetation water content. Estimated spatial covariance functions and soil moisture distributions are compared with those from Polarimetric Scanning Radiometer (PSR) soil moisture images taken during SMEX02.
H13C-0435 1340h
Estimation of Rootzone Soil Moisture and Land Surface Fluxes From Reference-Level Micrometeorology and Boundary Layer Observations
Rootzone soil moisture is the key land surface state due to its role in partitioning both the incoming radiation into latent and sensible heat fluxes and precipitation into infiltration, runoff, and subsequent evapotranspiration. These surface fluxes of moisture and energy are complex functions of uncertain atmospheric forcing and land surface characteristics (vegetation and soil properties). Because of the strong coupling between the land surface and overlying atmospheric boundary layer, valuable information about subsurface states and surface fluxes is contained in readily-available surface layer and boundary layer observations (micrometeorological temperature and humidity, radiosonde data, satellite-based soundings, etc.). In this study we apply the Ensemble Kalman Filter with a coupled land surface boundary layer model to assimilate micrometeorological and boundary layer observations to estimate rootzone soil moisture. Using the coupled model requires minimal auxiliary information and variables that are typically required as forcing for offline models can instead be assimilated, providing a further constraint on the flux estimates. The method is applied to the Central Facility region of the Southern Great Plains 1997 (SGP97) field experiment site. Synthetic experiments are performed to assess the performance of the assimilation method in both moisture-limited and energy-limited regimes and under uncertain and biased model parameters (soil hydraulic properties, vegetation characteristics, etc.) and forcing (radiation and precipitation). The ensemble approach is shown to not only provide reasonable estimates of mean rootzone soil moisture over time, but can also track its uncertainty, which is a complicated function of time as a result of model input uncertainty. By constraining the rootzone soil moisture, significant improvements in the surface turbulent fluxes are also seen. Results when assimilating the real SGP97 data also show significant improvements in rootzone soil moisture and surface flux estimates over simple forward modeling. The results from the simple coupled model used here indicate the potential for extension of this approach to large scale applications using more complex mesoscale models.
H13C-0436 1340h
Water Balance Study on a Semiarid Regional Landscape in South Texas: Effects of Anthropogenic Land Disturbance
While extensive vegetation manipulation has been encouraged by many administrators and extension groups, its effect on the water balance is complex and the hypothesis that removal of woody plants consistently reduces evapotranspiration, increases soil water content, and water yield remains unproven. This project focuses on examination of the effects of various land management practices on the overall water balance for semiarid regional landscapes. The project location is at the Wellhausen Ranch Research Station located near Laredo, TX, consisting of 5,280 acres of shrub landscape dominated by honey mesquite shrub species. This ranch has undergone various land disturbances such as root plowing and cattle overgrazing that have caused damage to the vegetation and natural communities. Five research sites were chosen within the ranch including a control site, a gravel dominated site, a root plowed site, an undisturbed site, and a second research site with different vegetative cover to represent different land use environments. Parameters that are being measured for the water balance study include precipitation, soil moisture, surface runoff, evaporation, and evapotranspiration. Preliminary results show that for the period of January to September of 2004, temperatures in the Wellhausen Ranch range from 29.1 °F to 106.9 °F, indicating hot summers and mild winters. 68 rainfall events have occurred, which resulted in 16.24 inches of total precipitation. Patterns were detected in soil moisture profiles reflecting the differences of soil moisture at different depths in the soil. Analysis of variance (ANOVA) indicates significant differences in the soil moisture in the five research sites. In addition, micro-lysimeter results show higher evaporation rates in the gravel dominated and the second research sites. These preliminary results indicate a potentially significant influence of anthropogenic land disturbance on a landscape water balance in the semiarid Nueces River basin.
H13C-0437 1340h
Evaluating Satellite Rainfall Estimates for Agro-hydrological Applications in Africa
Regional water balance techniques are used to monitor and forecast crop performance and flooding potentials around the world. In the last few years, satellite rainfall estimates (RFE) have become available at continental scales, which made it possible to develop operational regional water balance models for the monitoring of crops performance and flooding potentials in Africa and other regions of the world as part of an environmental early warning system . The accuracy of RFE in absolute terms and importantly as it relates to agricultural and hydrological applications have not been evaluated systematically. This study evaluated a subset of the Africa-wide RFE product by comparing station-rainfall data and RFE from 1996 to 2002 using over 100 rain-gauge stations from Ethiopia at a dekadal (~10-day) time step. The results showed a general under-estimation of RFE compared to station rainfall values. The correlation between station rainfall data and RFE varied highly from place to place and between seasons. On the other hand, the correlation improved significantly when comparison was made between RFE-derived crop water satisfaction index (WRSI) and station-rainfall-derived WRSI, indicating the usefulness of the RFE for agro-hydrological applications.
http://igskmncnwb015.cr.usgs.gov/adds/
H13C-0438 1340h
Comparison between Neural Network and Fuzzy Logic system for Soil Moisture Estimation using Microwave Remote Sensing Data
Artificial neural networks and Fuzzy logic have been applied to a wide range of problems in several disciplines. They have been successfully applied to image processing, and have shown a great potential in the classification of remote sensing data. However, a successful application of these methods in remote sensing data classification requires a good comprehension of the effect of their internal parameters and especially those that are related to the algorithm structure and to the training process. In this work we report the application of backpropagation neural network and fuzzy logic in estimating the soil moisture level using Synthetic Aperture Radar (SAR) data. The potential of SAR images in spatial soil moisture estimation depends on the ability of these algorithms to define the complex relationship that exists between the backscattered energy and the moisture content of the soil. A study area located in Oklahoma (97d35'W, 36d15'N) has been chosen for this project. Several textural measures derived from Radarsat-1 images acquired in Scansar Mode during the summer of 1997 were used as input for two algorithms. The soil moisture data measured by ESTAR Instrument (Electronically Scanned Thinned Array Radiometer) during the SGP97 campaign (operated by NASA) were used as truth data in the training process. The effect of some parameters related to the training process on classification performance was investigated for both methods. The preliminary results showed that for neural networks, the variations of the number of hidden layers and the number of nodes by layer have no significant effect on classification accuracy. However, the retained threshold value used in the output layer affects significantly the overall classification. Concerning, the fuzzy logic algorithm, the preliminary results showed that the cluster radius selection have a significant effect on classification accuracy.
H13C-0439 1340h
High-resolution Land Surface Modeling and Data Assimilation
A 51-year simulation of water and energy fluxes over the entire Arkansas-Red based was performed using a fully distributed land surface model. The simulations were performed at fine temporal (hourly) and spatial (1 squ. km) resolutions in an effort to bridge the gap between traditional hydrologic modeling, typically at final temporal spatial resolutions on relatively small catchment, and regional land surface modeling. Our approach in model validation is to focus on both the accuracy of streamflow simulations at the sub-basin scale and appropriate physically based description of heat and water exchange at the land surface-atmosphere interface because biases at the sub-basin scale may grow nonlinearly over time and lead to larger basin-scale biases. Preliminary analysis of the simulations shows that the spatial patterns of temporarily averaged water balance components are similar to published climatological patterns and clearly illustrates the strong east-west gradients of precipitation, runoff, and evapotranspiration. Streamflow accumulations at the sub-basin scale show good agreement betweens simulated and observed streamflow for catchments ranging in size between 880 km2 and 4211 km2. Results of energy flux simulations shown in this study are in reasonable agreement with observations. Analysis of the spatial distribution of precipitation and runoff highlights the similarities and differences between the two. The relationship between precipitation anomalies and runoff and soil water storage anomalies was examined. The study is ongoing and more validation at the sub-basin scale is being done. We hope that the results of this analysis will help clarify the sources of long-term hydrologic variability within the basin.
H13C-0440 1340h
Assessing Remotely Sensed Based Estimates of Evapotranspiration Using a Global Set of Evaluation Datasets.
Developing a globally robust algorithm for the prediction of surface heat fluxes is a significant challenge. Difficulties in capturing the hydro-climatic variability inherent at global scales have limited the extensive application of remote sensing approaches for characterizing surface heat flux behavior. An increased ability to capture the land surface variability has arisen with the development of a number of key remote sensing based products. Spatial and temporal fields of the surface temperature, land surface cover and vegetation distribution are globally available, offering increased ability to monitor hydrological patterns. Coupled with improved data availability is the growing availability of high quality, in-situ validation datasets, essential for the robust evaluation of model responses. Two such data sets are the WCRP/GEWEX/CEOP reference tower data and the FLUXNET tower data. The Coordinated Enhanced Observing Period (CEOP) activity is an element of the World Climate Research Program (WCRP), initiated by the Global Energy and Water Cycle Experiment (GEWEX). Along with FLUXNET, a global network of carbon dioxide and micro-meteorological tower sites, these programs provide measurements of water vapor and energy exchanges over diverse environments across the globe. Combined, the two datasets form a unique hydro-climatological database with global consistency over a range of climatic and vegetation conditions, making them ideally suited to robustly evaluate regional and meso-scale hydrological models and applications. In this study, observations from CEOP and FLUXNET are used to assess estimates of the evapotranspiration, determined using a number of approaches. The purpose of this analysis is to evaluate the adaptability of varied techniques to different climatic conditions and land cover types and conditions. Forcing data from validation tower sites and widely available remote sensing products are used to produce estimates of the land surface fluxes. Daily and 10-day averaged surface fluxes are computed and compared to in-situ observations. Monthly mean diurnal fluxes are also determined to assess the level of temporal variability throughout the observation period at each of the investigation sites. Comparisons show that model predictions of the energy fluxes indicate some promise towards the development of a robust algorithm to derive a global land surface evapotranspiration product.
H13C-0441 1340h
Spatially Explicit Observations to Elucidate Simple Scalars of Forest Canopy Transpiration Across Environmental Gradients
The ability to scale from point measurements to watersheds has been a key goal of hydrology. Assumptions are often made that averaging point measurements and scaling them up using a cookie-cutter or paint-by-numbers approach will capture relevant spatial gradients. To test this, we chose a site in the Chequamegon National Forest near Park Falls, WI because of its proximity to the WLEF Ameriflux tower providing kilometer scale estimates of water fluxes from a heterogeneous forest. We used a cyclic sampling design for all 144 plots of spatial measurements within a 1.5 ha area, in order to efficiently quantify spatial trends using geostatistics. Spatial data was collected for sap flux using Granier type sensors daily for ten days in 170 trees representing 7 species, including aspen, alder, and white cedar. Aspen is a dominant species in the managed forests around the WLEF tower and we have previously shown it to have the highest transpiration rates per unit leaf area of all dominant species in the area. Consequently, for this study we focused on aspen. Spatial soil moisture, vapor pressure deficit, and leaf area index were also measured periodically at the same 144 plots. We found that the semivariagram of soil moisture showed a range of 110 meters on a low soil moisture day and 80 meters on a high soil moisture day. When we quantified sap flux per unit xylem area across a 105-meter long gradient from a wetland to an upland we found no differences. However, once we scaled the sap flux measurements to the whole tree using basal area, there was more than a 100 percent increase in whole tree water use in the upland area in comparison to the wetland area. Thus, we will test the hypothesis that in the absence of moisture stress, canopy transpiration in aspen varies spatially with allometrically scaled sapwood area and leaf area and not as a function of sap flux per unit sapwood area.
H13C-0442 1340h
Elevation controls on timing and quantity of water yield from a semi-arid catchment
Water supply in many semi-arid regions is derived from mountain precipitation, which varies in depth and phase with elevation. Predicting water yield from mountainous regions depends on the spatial and temporal distributions of the source area precipitation, and energy balance, which controls both snowmelt and evapotranspiration. The Simultaneous Heat and Water (SHAW) model was developed to simulate the response of the water balance at the component level to weather and soil. The SHAW model was applied to 4 years of weather station and soil moisture and temperature data at two elevations (1610.5 m and 1142.4 m) in Dry Creek watershed near Boise, ID to determine the annual variability in elevational controls on water yield. In particular, we were interested in improving our ability to better predict water yield for low snowpack conditions. Precipitation fell primarily between October and April at both elevations. The average annual precipitation was 57 cm and 33 cm at the upper and lower sites respectively. The maximum daytime air temperature at the lower site was generally higher than at the upper site but the nighttime temperatures were similar. Evapotranspiration was greater at the upper site, which remained wetter longer, due to the greater precipitation. No surface runoff was measured over the four year period. Water yield was equated with deep percolation from the soil column which was seasonal, and occurred from January through March at both sites. The onset and duration of deep percolation at both sites was dependent on the timing of increasing air temperature and extent of spring precipitation. Whereas total winter precipitation was found to be a first-order control on water yield, timing of spring rains was found to be an important second order control on deep percolation and runoff generation.
H13C-0443 1340h
GRAIN SIZE DEPENDENT, SHORT TIME SCALE WATERSHED TERRAIN EVOLUTION MODEL USING A PATH SAMPLING MONTE CARLO METHOD
We present a new GRASS GIS module r.terradyn that evolves a given terrain over short time scales using sediment flux information provided by the SIMWE (SImulated Water Erosion) GRASS GIS modules r.sim.water and r.sim.sediment originally developed by Mitas and Mitasova (1998). SIMWE is a distributed, bivariate, steady-state watershed scale sediment erosion, transport, and deposition model that employs a path sampling Monte Carlo method in which erosion, transport, and deposition conditions are treated as a continuous field, resulting in fully distributed erosion/deposition patterns. Module r.terradyn modifies the original digital elevation model (DEM) per rainfall event, which is then used as the input DEM for subsequent SIMWE and r.terradyn iterations. New techniques were derived that include the application of a gravitational diffusion term, an approximate Neumann boundary condition routine for use with GRASS GIS module r.slope.aspect, a comparative band-pass filter for numerical stability of the iterative feedback system, and a simple rainfall excess calculation methodology derived from accumulated runoff curve number tables that enables spatially distributed infiltration. Application of r.terradyn to a sample watershed demonstrates results for distributed land cover and infiltration and for various grain sizes. Terrain change impact from a disturbed area is also presented. Preliminary comparisons to field observations and total discharge data are currently being used to calibrate model parameters. Verification of the model is still ongoing as data becomes available. The influence of grain size dependent transport mechanisms on short-term and long-term topological changes induced by human impact, such as mining and construction, may lead to the determination of the optimum location, size, and frequency of control measures to more cost effectively meet emerging TMDL requirements.
H13C-0444 1340h
Identification and Characterization of Potential Ground Targets for Satellite Microwave Sensor Calibration Monitoring
AMSR-E (on Aqua) was launched in May 2002, one of a series of microwave radiometers on Earth-orbiting satellites that also includes SMMR, SSM/I, TMI and AMSR (on ADEOS-II)and Windsat. These instruments all have onboard calibration systems designed to enable calibrated brightness temperatures to be derived in the ground processing. However, all satellite radiometer data require fine-tuning in the post-launch overall system calibration to remove residual radiometer and antenna system calibration artifacts. These can arise from a variety of causes such as uncorrected attitude errors, instrument misalignment or pointing errors, instrument thermal gradients, reflector emissivity errors, calibration coefficient errors, component degradation, etc. Post-launch calibration fine-tuning usually involves finding suitable external targets, i.e., homogeneous Earth surface areas of large spatial extent, such as to be unaffected by nonlinear averaging or antenna pattern effects. The brightness temperature characteristics of the external targets should be stable and well-known to be used as reliable absolute references. Calm ocean areas have been used as cold reference targets to cross-calibrate the SSM/I, TMI and AMSR-E sensors at the cold end (80-150 K) of the typical Earth-view brightness temperature dynamic range. However, even if the cold end of the calibration is adjusted accurately, such that ocean geophysical products can be accurately derived, the warm end of the calibration may still be inaccurate, giving rise to inaccuracies in geophysical retrievals over land. There is less knowledge of what constitutes a suitable target at the mid-range (~200 K) or warm end (250-300 K) of the scale. Ice sheets have been considered for the mid-range, and tropical forests for the warm end, but there has been relatively little analysis of the radiative transfer modeling accuracy, and spatial and temporal stability, of the brightness temperatures of ice sheet and forest locations that might make them suitable choices as external targets for spaceborne radiometer system calibration. We will present a systematic investigation of the brightness temperature spatial and temporal variability of a range of ice-sheet and forest sites in an effort to characterize and select optimum sites for long-term spaceborne radiometer system calibration monitoring over land. Such sites may be intstrumented and maintained as calibration sites for future missions such as CMIS and Hydros.
H13C-0445 1340h
Estimating Leaf Area Index in Forests Using Airborne Laser Swath Mapping Data
In recent years airborne laser swath mapping (ALSM) has enabled topographic mapping at the several centimeter scale with meter scale horizontal sampling. ALSM has made it possible, for the first time, to study three-dimensional foliage structure on spatial scales extending from meters to tens of kilometers in a consistent geodetic frame of reference. A multiscale filter has been developed to separate ALSM returns from ground and vegetation over forests using statistical decision theory methods. As a result, it is possible to calculate probabilities of laser light interception by the canopy and by the ground, enabling a direct estimation of leaf area index (LAI) as a function of sample density. Most ALSM systems employ small laser footprints (less than 1 m diameter) to approximate point-to-point range measurements. As a result, individual tree canopies can often appear effectively opaque to a single laser shot. The ground is only detected when some fraction of the transmitted laser pulses pass through naturally occurring gaps in the canopy and intercept the surface (known as ground shots). LAI is an important parameter for understanding forest hydrologic processes because it impacts evaporation and transpiration rates, as well as the interception of precipitation. As with attempts to estimate LAI using microwave and multispectral data, ALSM-based estimates can suffer from saturation when extremely dense tree crowns occlude lower canopy layers. However, the high-resolution three-dimensional ALSM measurements allow for improved modeling of the discrete structure of the forest canopy, which makes it possible to characterize the saturation phenomenon. Simulated ALSM data and ground truthing are used to characterize the accuracy of LAI estimates from ALSM data acquired over mixed deciduous/coniferous forests of the Southeastern United States and to determine the estimation error caused by the saturation phenomenon as a function of canopy parameters.
H13C-0446 1340h
Hydrodynamic Modeling Approaches for Surface Water Reservoirs
Hydrologic modeling of surface water reservoirs is an effective tool in evaluating different water management options for addressing regional water issues in Florida. However, modeling reservoir water dynamics could be challenging because of the difference in scale between canals and the entire reservoir. Water pumped into the reservoirs is first discharged into canals inside the reservoirs, which distributes the water. The canal eventually overflows and water floods all the reservoir. Two modeling approaches to simulate this process were tested on two reservoirs using the integrated MIKE-SHE and MIKE 11 model. The first approach simulates the 1- D flow in the canal in a link-node model and once water floods, it is modeled as 2-D flow. The second approach simulates the entire impoundment as a canal. In both reservoirs, Modeling Approach 1 resulted in overestimation of peaks and poor results. Modeling Approach 2 showed considerable improvements in the results and a satisfactory match between observed and simulated water levels. The difference is attributed to the difficulty in representing the canal flooding process in hydrodynamic models.
H13C-0447 1340h
Contradictions in defining land use change and impervious surface area using satellite
I will discuss two aspects of land surface analysis in terms of inherent contradiction in their definition. The first pertains to how one analyzes consistent temporal changes in land use using satellite imagery at the pixel level. The second pertains to the definition of impervious surface area, alternately from a surface and a satellite perspective.
H13C-0448 1340h
HYDROLOGICAL MODELLING OF CHERIAL WATERSHED INTEGRATING REMOTE SENSING AND GEOGRAPHICAL INFORMATION SYSTEM (GIS)
The increasing population growth is continuing to exert extra pressure on existing water resources all over the world. An imperative need for the development and judicious use of these resources is therefore essential. Rainfall in southern India is very erratic, unpredictable, uneven and distributed over a short period of 3-4 Months. Out of 4000 billion cubic meters of rainfall received annually, 41% is lost as evaporation and transpiration, 40% lost as runoff into seas and 10% seeps in for recharging groundwater. As a result Krishna and Godavari rivers of southern India are almost in dry conditions throughout the year with farmers suffering from droughts since past 20 years especially in the state of Andhra Pradesh. An imperative need for the development and judicious use of these resources is therefore essential for conservation of water resources and maintaining the hydrologic table when water is becoming a scarce material. Keeping this in view an integrated model is developed for the chronically drought prone area of Cherial watershed in Warangal district of Andhra Pradesh using Remote sensing and GIS techniques. This model explores and suggests cost-effective and sustainable methods of increasing the crop yield by increasing the ground water potential artificially. The main objective of the study is to evaluate both surface and groundwater resources in the region and develop methods for its efficient utilization and sustainable management. Remote sensing and GIS applications are adopted as an effective tool in meeting the objective of the study. The thematic layers v.i.z. drainage pattern, land use/ land cover, hydrogeomorphology, slope, soil, physiography and ground water prospects are all derived from IRS-ID PAN + LISS-III merged satellite imagery and Survey of India (SOI) topomaps using visual interpretation technique. These maps are then converted to digital format using AutoCAD software and further integrated using Arc/Info and ArcView GIS software for the generation of final action plan maps for water and land resources development which are optimally suitable to the terrain and to the development of water resources so that the level of production is sustained without decline over time. Effective soil and water conservation measures such as construction of recharge structures are recommended to increase the subsurface aquifer capacity. Suitable cropping patterns duly considering the climatic and soil moisture conditions are suggested which help in reduced soil erosion, increased moisture conservation and improved productivity of the soil. The hydrological model developed aims for optimum development of water resources required to meet basic minimum needs of farmers thereby improving their socio-economic conditions and helps in evolving a broad national policy which can be applied by decision makers for sustainable development of any given watershed area.
H13C-0449 1340h
Wenatchee River, Washington, Water Temperature Modeling and Assessment Using Remotely Sensed Thermal Infrared and Instream Recorded Data
The stream water spatial and temporal temperature patterns of the Wenatchee River, WA are assessed based on temperature data recorded by instream data loggers in the dry season of 2002 and thermal infrared imagery from August 16th 2002. To gain insights into the possible thermal behavior of the river, the stream temperature model Qual2K (Chapra and Pelletier, 2003) is extended beyond its calibration (10-16 August 2002) and confirmation (9-11 September 2002) periods for use with different meteorological, shade and flow conditions. The temperature longitudinal profile of the Wenatchee River is influenced by the temperature regime in Lake Wenatchee, the source of the Wenatchee River. Model simulations performed at 7-day average with 2-year return period flow conditions show that the potential (maximum average across all reaches) temperature (the temperature that would occur under natural conditions) is about 19.8 deg. C. For the 7-day average with 10-year return period flow conditions the potential temperature increases to about 21.2 deg. C. The simulation results show that under normal flow and meteorological conditions the water temperature exceeds the current water quality standards. Model simulations performed under the 7-day average with 10-year return period flow conditions and a climate change scenario show that the average potential temperature across all reaches can increase by as much as 1.3 deg. C compared to the case where climate change impact is not taken into account. Thermal infrared (TIR) derived stream temperature data were useful for describing spatial distribution patterns of the Wenatchee River water temperature. The TIR and visible band images are effective tools to map cold water refugia for fish and to detect regions that can be improved for fish survival. The images collected during the TIR survey and the TIR derived stream temperature longitudinal profile helps pinpoint additional instream monitoring locations that avoid regions of backwater, cool or warm pockets or regions affected by tributary influence, that are inappropriate for stream temperature monitoring. Groundwater input is difficult to detect from the TIR images in the case of a relatively large river such the Wenatchee River.
H13C-0450 1340h
Evaluation of Temporal and Spatial Distribution of Error in Modeled Evapotranspiration Estimates
Evapotranspiration (ET) constitutes a significant portion of Florida's water budget, and is second only to rainfall. Consequently, accurate ET estimates are very important for hydrologic modeling work. However, in comparison to rainfall, relatively few ground stations exist for the measurement of this important model input. Consequently, ET estimates produced by models are often subject to error. Satellite-based ET estimates provide an unprecedented opportunity to measure actual ET in sparsely monitored watersheds. They also provide a basis for comparing errors in modeled actual ET estimates that are induced due to the following reasons: 1) spatial interpolation and data-filling methods; 2) inaccurate and sparse meteorological data; and, 3) simplified parameterization schemes. In this study, satellite-based daily actual ET estimates from the Water Conservation Area 3 (WCA-3) watershed in South Florida, USA, are compared with those obtained from a calibrated finite-volume regional hydrologic model for the 1998 and 1999 calendar years. The satellite-based ET estimates used in this study compared well with measured ground-based actual ET data. The WCA-3 watershed is an integral part of Florida's remnant Everglades, and covers an area of approximately 2,400 square kilometers. It is compartmentalized by several levees and road embankments, and drained by several major canals. It also serves as a major habitat for many wildlife species, a source for urban water supply and an emergency storage area for flood water. The WCA-3 is located east of the Big Cypress National Preserve, and north of the Everglades National Park. Despite its significance, WCA-3 has relatively few ET monitoring stations and meteorological stations. Consequently, it is ideally suited for evaluating and quantifying errors in simulated actual ET estimates. The Regional Simulation Model (RSM) developed by the South Florida Water Management District is used for the modeling of these ET estimates. The RSM is an implicit, finite-volume, continuous, distributed, integrated surface/ground-water model, capable of simulating one-dimensional canal/stream flow and two-dimensional overland flow in arbitrarily shaped areas using a variable triangular mesh. The RSM has several options for modeling actual ET. An empirical parameterization scheme that is dependent on land-cover, water-depth and potential ET is used in this study for estimating actual ET. The parameter-sensitivities of this scheme are investigated and analyzed for several predominant land-cover classes, and dry- and wet-soil conditions. The RSM is calibrated and verified using historical time-series data from 1988 to 1995, and 1996 to 2000, respectively. All sensitivity and error analyses are conducted using estimates from the verification period.
H13C-0451 1340h
An evaluation of complementary relationship assumptions
Complementary relationship (CR) models, based on Bouchet's (1963) somewhat heuristic CR hypothesis, are advantageous in their sole reliance on readily available climatological data. While Bouchet's CR hypothesis requires a number of questionable assumptions, CR models have been evaluated on variable time and length scales with relative success. Bouchet's hypothesis is grounded on the assumption that a change in potential evapotranspiration (E$_p}$) is equal and opposite in sign to a change in actual evapotranspiration (E$_{a}$), i.e., $-$dE$_{p}$ $/$ dE$_{a}$ $=$ 1. In his mathematical rationalization of the CR, Morton (1965) similarly assumes that a change in potential sensible heat flux (H$_{p}$) is equal and opposite in sign to a change in actual sensible heat flux (H$_{a}$), i.e., $-$dH$_{p}$ $/$ dH$_{a}$ $=$ 1. CR models have maintained these assumptions while focusing on defining E$_{p}$ and equilibrium evapotranspiration (E$_{po}$). We question Bouchet and Morton's aforementioned assumptions by revisiting CR derivation in light of a proposed variable, $\phi$ $=$ $-$dE$_{p}$$/$dE$_{a}$. We evaluate $\phi$ in a simplified Monin Obukhov surface similarity framework and demonstrate how previous error in the application of CR models may be explained in part by previous assumptions that $\phi$=1. Finally, we discuss the various time and length scales to which $\phi$ may be evaluated.
H13C-0452 1340h
An Estimation of Atmospheric Moisture Transport Over Indian Sub-continent by Using 4D-VAR Data Assimilation System
By using an atmospheric general circulation model and a 4-dimensional variational (4D-VAR) data assimilation system, origin and transport processes of water over Indian sub-continent in summer and autumn are examined. Koster et al.(2004) pointed out that Indian sub-continent is one of the specific locations for which soil moisture anomalies have a substantial impact on precipitation in boreal summer (JJA) through a multimodel experiment of land-atmosphere coupling strength. On the other hand, the diagnostic estimation of recycling ratio by Trenberth (1999), which is defined as the relative contribution of the locally evaporated water to the precipitating water in the same region, shows that the value of recycling ratio over Indian sub-continent in autumn is higher than that in summer. Then we investigate the seasonal march of atmospheric moisture transport of precipitating water over Indian sub-continent, especially focus on the differences between boreal summer and autumn season. The global distribution of recycling ratio in our model has been well-reproduced in comparison with Trenberth_fs results. The moisture source distribution of precipitating water over India can be detected through sensitivity experiments using our 4D-VAR data assimilation system. These calculations are conducted on the Earth Simulator. In summer monsoon season, two typical cases can be seen in our model. One is that the atmospheric moisture which is a source of rainfall over Indian subcontinent is mainly supplied from Arabian Sea transported by Somali jet, which results in the typical summer Indian monsoon circulation. In addition, some other cases show that the atmospheric moisture is mainly supplied by the evapotranspiration from Indian land area, that is, locally recycled water. These results indicate that the amount of water supply over India is mainly determined by the combination of these two cases. And the local recycling of water is considered to be important to decide the summer precipitation over India as well as the transport of water vapor from the outside of Indian sub-continent. In autumn, the contribution of atmospheric moisture transport from the outside becomes relatively smaller as accompanied by summer monsoon withdrawal. As the result, water recycling in this area is more important than that in summer. This implies that over India the condition of soil moisture through summer monsoon season could strongly affect the rainfall activity in autumn. We are now developing the 4D-VAR data assimilation system using a coupled atmosphere-ocean-land global circulation model, and this method will be applied to the interannual cases in the near future.
H13C-0453 1340h
BAYESIAN NETWORK STRUCTURE LEARNING FOR URBAN LAND USE CLASSIFICATION FROM LANDSAT ETM+ AND ANCILLARY DATA
Recognizing urban information from the satellite imagery is problematic due to the diverse features and dynamic changes of urban landuse. The use of Landsat imagery for urban land use classification involves inherent uncertainty due to its spatial resolution and the low separability among land uses. To resolve the uncertainty problem, we investigated the performance of Bayesian networks to classify urban land use since Bayesian networks provide a quantitative way of handling uncertainty and have been successfully used in many areas. In this study, we developed the optimized networks for urban land use classification from Landsat ETM+ images of Marina del Rey area based on USGS land cover/use classification level III. The networks started from a tree structure based on mutual information between variables and added the links to improve accuracy. This methodology offers several advantages: (1) The network structure shows the dependency relationships between variables. The class node value can be predicted even with particular band information missing due to sensor system error. The missing information can be inferred from other dependent bands. (2) The network structure provides information of variables that are important for the classification, which is not available from conventional classification methods such as neural networks and maximum likelihood classification. In our case, for example, bands 1, 5 and 6 are the most important inputs in determining the land use of each pixel. (3) The networks can be reduced with those input variables important for classification. This minimizes the problem without considering all possible variables. We also examined the effect of incorporating ancillary data: geospatial information such as X and Y coordinate values of each pixel and DEM data, and vegetation indices such as NDVI and Tasseled Cap transformation. The results showed that the locational information improved overall accuracy (81%) and kappa coefficient (76%), and lowered the omission and commission errors compared with using only spectral data (accuracy 71%, kappa coefficient 62%). Incorporating DEM data did not significantly improve overall accuracy (74%) and kappa coefficient (66%) but lowered the omission and commission errors. Incorporating NDVI did not much improve the overall accuracy (72%) and k coefficient (65%). Including Tasseled Cap transformation reduced the accuracy (accuracy 70%, kappa 61%). Therefore, additional information from the DEM and vegetation indices was not useful as locational ancillary data.
H13C-0454 1340h
Radar Satellite (InSAR) Assessment of Hydrodynamics Near the All-American Canal (Calexico/Mexicali Region, Rio Colorado)
Dispute settlement over groundwater issues is hampered by the fact that groundwater is not discussed in existing bilateral treaties between Mexico and the United States whereas aquifers frequently span the border zone. Accurate, international data on groundwater resources and dynamics are therefore needed to assist formulation of bi-national groundwater policy, particularly in the Colorado River delta region. We will present our preliminary InSAR (Interferometric Synthetic Aperture Radar) measurements of cm-scale vertical displacements of the surface above the Mexicali Valley aquifer/ Mesa d'Andrade, using these as a proxy for aquifer depletion and recharge events. It is anticipated that this InSAR monitoring will allow characterization of aquifer behavior across the border zone over the past decade, and prior to (or during) the planned lining of the All-American Canal and the federal reductions in Colorado River surface water allocations to urban and/or rural California consumers. Either action could seriously alter a major aquifer recharge zone and, consequently, groundwater volumes in Mexico (655 private and federal pumping sites) and southeastern California. We will present preliminary deformation maps for a roughly 650 km2 area of the Mexicali-Calexico region; a first step in characterizing regional and local `pre-lining' subsidence signals due to groundwater pumping, geothermal energy operations, tectonic creep and, possibly, changes in soil properties.