Arkansas-Red River Basin Research Articles

Dual state/rainfall correction via soil moisture assimilation for improved streamflow simulation: evaluation of a large-scale implementation with Soil Moisture Active Passive (SMAP) satellite data

Abstract. Soil moisture (SM) measurements contain information about both pre-storm hydrologic states and within-storm rainfall estimates, both of which are required inputs for event-based streamflow simulations. In this study, an existing dual state/rainfall correction system is extended and implemented in the 605 000 km2 Arkansas–Red River basin with a semi-distributed land surface model. The Soil Moisture Active Passive (SMAP) satellite surface SM retrievals are assimilated to simultaneously correct antecedent SM states in the model and rainfall estimates from the Global Precipitation Measurement (GPM) mission. While the GPM rainfall is corrected slightly to moderately, especially for larger events, the correction is smaller than that reported in past studies due primarily to the improved baseline quality of the new GPM satellite product. In addition, rainfall correction is poorer in regions with dense biomass due to lower SMAP quality. Nevertheless, SMAP-based dual state/rainfall correction is shown to generally improve streamflow estimates, as shown by comparisons with streamflow observations across eight Arkansas–Red River sub-basins. However, more substantial streamflow correction is limited by significant systematic errors present in model-based streamflow estimates that are uncorrectable via standard data assimilation techniques aimed solely at zero-mean random errors. These findings suggest that more substantial streamflow correction will likely require better quality SM observations as well as future research efforts aimed at reducing systematic errors in hydrologic systems.

A Framework for Diagnosing Factors Degrading the Streamflow Performance of a Soil Moisture Data Assimilation System

Abstract Data assimilation (DA) techniques have been widely applied to assimilate satellite-based soil moisture (SM) measurements into hydrologic models to improve streamflow simulations. However, past studies have reached mixed conclusions regarding the degree of runoff improvement achieved via SM state updating. In this study, a synthetic diagnostic framework is designed to 1) decompose the random error components in a hydrologic simulation, 2) quantify the error terms that originate from SM states, and 3) assess the effectiveness of SM DA to correct these random errors. The general framework is illustrated through a case study in which surface Soil Moisture Active Passive (SMAP) data are assimilated into a large-scale land surface model in the Arkansas–Red River basin. The case study includes systematic error in the simulated streamflow that imposes a first-order limit on DA performance. In addition, about 60% of the random runoff error originates directly from rainfall and cannot be corrected by SM DA. In particular, fast-response runoff dominates in much of the basin but is relatively unresponsive to state updating. Slow-response runoff is strongly controlled by the bottom-layer SM and therefore only modestly improved via the assimilation of surface measurements. Combined, the total runoff improvement in the synthetic analysis is small (<10% over the basin). Improvements in the real SMAP-assimilated case are further limited due to systematic error and other factors such as inaccurate error assumptions and SMAP rescaling. Findings from the diagnostic framework suggest that SM DA alone is insufficient to substantially improve streamflow estimates in large basins.

Evaluation of TOPMODEL-Based Land Surface–Atmosphere Transfer Scheme (TOPLATS) through a Soil Moisture Simulation

Abstract Better quantification of the spatiotemporal distribution of soil moisture across different spatial scales contributes significantly to the understanding of land surface processes on the Earth as an integrated system. While observational data for root-zone soil moisture (RZSM) often have sparse spatial coverage, model-simulated soil moisture may provide a useful alternative. TOPMODEL-Based Land Surface–Atmosphere Transfer Scheme (TOPLATS) has been widely studied and actively modified in recent years, while a detailed regional application with evaluation currently is still lacking. Thus, TOPLATS was used to generate high-resolution (30 arc s) RZSM based on coarse-scale (0.125°) forcing data over part of the Arkansas–Red River basin. First, the simulated RZSM was resampled to coarse scale to compare with the results of Mosaic, Noah, and VIC from NLDAS. Second, TOPLATS performance was assessed based on the spatial absolute difference among the models. The comparison shows that TOPLATS performance is similar to VIC, but different from Mosaic and Noah. Last, the simulated RZSM was compared with in situ observations of 16 stations in the study area. The results suggest that the simulated spatial distribution of RZSM is largely consistent with the distribution of topographic index (TI) in most instances, as topography was traditionally considered a major, but not the only, factor in horizontal redistribution of soil moisture. In addition, the finer-resolution RZSM can reflect the in situ soil moisture change at most local sites to a certain degree. The evaluation confirms that TOPLATS is a useful tool to estimate high-resolution soil moisture and has great potential to provide regional soil moisture estimates.

Analysis of Precipitation Projections over the Climate Gradient of the Arkansas Red River Basin

AbstractIncreases in the frequency and intensity of extreme precipitation are projected for most U.S. regions under climate change. There is a high degree of uncertainty, however, in precipitation regime changes across the large precipitation gradient of the Arkansas Red River basin (ARRB). The authors analyzed future precipitation regimes using two statistical downscaling datasets based on the scenarios from phase 5 of the Coupled Model Intercomparison Project (CMIP5). Seasonal precipitation in low-to-high quantiles was calculated and compared for the southern ARRB where the downscaled data were available. The results showed a generally comparable shift in precipitation patterns and amounts between the two datasets. However, some spatial variation of precipitation amount change exists, and the direction of change could be opposite for the summer. Both datasets showed that the top 10% of monthly precipitation amounts could increase for the southern ARRB, mostly ranging from 5–10 mm month−1 for the early part of the century (2010–39) to 15–30 mm month−1 for the midcentury (2040–69) as compared with the historical period (1968–97). The maximum monthly precipitation could increase by up to 150 mm in both datasets by the midcentury. Precipitation was projected to increase regardless of quantile during both winter and spring but tended to decrease during summer and autumn. More-frequent and higher-intensity rainfall events were expected for the eastern part of the basin, and longer and drier dry periods were expected for the western basin. Conservation strategies and sustainable water management should consider the regional differences in the projected changes in precipitation regimes for the basin under climate change.

Atmospheric Contributors to Heavy Rainfall Events in the Arkansas-Red River Basin

This study analyzed the top 1% 24-hour rainfall events from 1994 to 2013 at eight climatological sites that represent the east to west precipitation gradient across the Arkansas-Red River Basin in North America. A total of 131 cases were identified and subsequently classified on the synoptic-scale, mesoscale, and local-scale to compile a climatological analysis of these extreme, heavy rainfall events based on atmospheric forcings. For each location, the prominent midtropospheric pattern, mesoscale feature, and predetermined thermodynamic variables were used to classify each 1% rainfall event. Individual events were then compared with other cases throughout the basin. The most profound results were that the magnitudes of the thermodynamic variables such as convective available potential energy and precipitable water values were poor predictors of the amount of rainfall produced in these extreme events. Further, the mesoscale forcings had more of an impact during the warm season and for the westernmost locations, whereas synoptic forcings were extremely prevalent during the cold season at the easternmost locations in the basin. The implications of this research are aimed at improving the forecasting of heavy precipitation at individual weather forecasts offices within the basin through the identified patterns at various scales.

An Observing System Simulation Experiment for the Aquarius/SAC-D Soil Moisture Product

An Observing System Simulation Experiment (OSSE) for the Aquarius/SAC-D mission has been developed for assessing the accuracy of soil moisture retrievals from passive L-band remote sensing. The implementation of the OSSE is based on the following: a 1-km land surface model over the Red-Arkansas River Basin, a forward microwave emission model to simulate the radiometer observations, a realistic orbital and sensor model to resample the measurements mimicking Aquarius operation, and an inverse soil moisture retrieval model. The simulation implements a zero-order radiative transfer model. Retrieval is performed by direct inversion of the forward model. The Aquarius OSSE attempts to capture the influence of various error sources, such as land surface heterogeneity, instrument noise, and retrieval ancillary parameter uncertainty, all on the accuracy of Aquarius surface soil moisture retrievals. In order to assess the impact of these error sources on the estimated volumetric soil moisture, a quantitative error analysis is performed by comparison of footprint-scale synthetic soil moisture with “true” soil moisture fields obtained from the direct aggregation of the original 1-km soil moisture field input to the forward model. Results show that, in heavily vegetated areas, soil moisture retrievals have a positive bias that can be suppressed with an alternative aggregation strategy for ancillary parameter vegetation water content (VWC). Retrieval accuracy was also evaluated when adding errors to 1-km VWC (which are intended to account for errors in VWC derived from remote sensing data). For soil moisture retrieval root-mean-square error on the order of 0.05 m <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> /m <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> , the error in VWC should be less than 12%.

Estimation of Satellite Rainfall Error Variance Using Readily Available Geophysical Features

The present study addresses the estimation of error variance (mean square error, MSE) of three satellite rainfall products: i) Tropical Rainfall Measuring Mission (TRMM) Multi-satellite Precipitation Analysis (TMPA) product of 3B42RT; ii) Climate Prediction Center (CPC) Morph (CMORPH); and iii) Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks-Cloud Classification System (PERSIANN-CCS). Nonlinear regression model is used to fit the response variable (satellite rainfall error variance) with explanatory variable (satellite rainfall rate) by grouping them as function of three key geophysical features: topography, climate, and season. The results of the study suggest that the error variance of a rainfall product is strongly correlated with rainfall rate and can be expressed as a power-law function. The geophysical feature based error classification analysis helps in achieving superior accuracy for prognostic error variance quantification in the absence of ground truth data. The multiple correlation coefficients between the estimated and observed error variance over an independent validation region (Upper Mississippi River basin) and time period (2007-2010) are found to be 0.75, 0.86, and 0.87 for 3B42RT, CMORPH, and PERSIANN-CCS products, respectively. In another validation region (Arkansas-Red River basin), the correlation coefficients are 0.59, 0.89, and 0.92 for the same products, respectively. Results of the assessment of error variance models reveal that the type of error component present in a satellite rainfall product directly impacts the accuracy of estimated error variance. The model estimates the error variance more accurately when the precipitation error components are mostly hit bias or false precipitation, while for a product with extensive missed precipitation, the accuracy of estimated error variance is significantly compromised. The study clearly demonstrates the feasibility of quantifying the error variance of satellite rainfall products in a spatially and temporally varying manner using readily available geophysical features and rainfall rate. The study is a path finder to a globally applicable and operationally feasible methodology for error variance estimation at high spatial and temporal scales for advancing satellite rainfall applications in ungauged basins.

Evaluation of Operational National Weather Service Gridded Flash Flood Guidance over the Arkansas Red River Basin

AbstractNational Oceanic and Atmospheric Administration's National Weather Service (NWS) flash flood warnings are issued by Weather Forecast Offices and are underpinned by information from the Flash Flood Guidance (FFG) system operated by the River Forecast Centers (RFCs). This study focuses on the quantitative evaluation and limitations of the FFG system using reported flash flood cases in 2010 and 2011. The flash flood reports were obtained from the NWS Storm Event database for the Arkansas‐Red Basin RFC (ABRFC). The current FFG system at the ABRFC provides gridded flash flood guidance (GFFG) system using the NWS Hydrology Laboratory‐Research Distributed Hydrologic Model to translate the upper zone soil moisture to estimates of Soil Conservation Service Curve Numbers. Comparisons of the GFFG and real‐time Multisensor Precipitation Estimator‐derived Quantitative Precipitation Estimate for the same duration and location were used to analyze the success of the system. Typically, the six‐hour duration was characterized by higher probability of detection values than the three‐hour duration, which highlights the difficulty of hydrologic process estimation for shorter time scales. The current system does not take into account physical characteristics such as land use, including irrigated agricultural farm and urban areas, hence, overly dry soil moisture estimates over these areas can lower the success rate of the GFFG product.

Climatological Drought Analyses and Projection Using SPI and PDSI: Case Study of the Arkansas Red River Basin

This paper examines past drought and assesses future drought scenarios for the Arkansas Red River Basin using two common drought indexes, the Standardized Precipitation Index (SPI) and Palmer Drought Severity Index (PDSI). Historical climate data within the 1900-2009 time frame were used to derive the past drought index estimates. The projected climate data under two greenhouse gas emission scenarios from 16 global climate models (GCMs) after bias correction and statistical downscaling were applied in drought occurrence fre- quency and affected area projection. The results derived from the SPI and PDSI show that widespread droughts mainly took place in the 1910s, 1930s, 1950s, and 1960s in the Arkansas Red River Basin, which agrees well with the historical climate record. Both the SPI and PDSI project that more frequent and severe droughts will appear in the second part of the 21st century under both of the emissions scenarios. Future PDSI projects that more severe droughts will occur in the western parts of this basin under one scenario. DOI: 10.1061/(ASCE)HE.1943- 5584.0000619. © 2013 American Society of Civil Engineers. CE Database subject headings: Arkansas; River basins; Climate change; Droughts; Case studies. Author keywords: Arkansas Red River Basin; Drought; Climate change; Projection; Global climate models (GCMs).

Short-term ensemble streamflow forecasting using operationally-produced single-valued streamflow forecasts – A Hydrologic Model Output Statistics (HMOS) approach

Summary We present a statistical procedure for generating short-term ensemble streamflow forecasts from single-valued, or deterministic, streamflow forecasts produced operationally by the U.S. National Weather Service (NWS) River Forecast Centers (RFCs). The resulting ensemble streamflow forecast provides an estimate of the predictive uncertainty associated with the single-valued forecast to support risk-based decision making by the forecasters and by the users of the forecast products, such as emergency managers. Forced by single-valued quantitative precipitation and temperature forecasts (QPF, QTF), the single-valued streamflow forecasts are produced at a 6-h time step nominally out to 5 days into the future. The single-valued streamflow forecasts reflect various run-time modifications, or “manual data assimilation”, applied by the human forecasters in an attempt to reduce error from various sources in the end-to-end forecast process. The proposed procedure generates ensemble traces of streamflow from a parsimonious approximation of the conditional multivariate probability distribution of future streamflow given the single-valued streamflow forecast, QPF, and the most recent streamflow observation. For parameter estimation and evaluation, we used a multiyear archive of the single-valued river stage forecast produced operationally by the NWS Arkansas-Red River Basin River Forecast Center (ABRFC) in Tulsa, Oklahoma. As a by-product of parameter estimation, the procedure provides a categorical assessment of the effective lead time of the operational hydrologic forecasts for different QPF and forecast flow conditions. To evaluate the procedure, we carried out hindcasting experiments in dependent and cross-validation modes. The results indicate that the short-term streamflow ensemble hindcasts generated from the procedure are generally reliable within the effective lead time of the single-valued forecasts and well capture the skill of the single-valued forecasts. For smaller basins, however, the effective lead time is significantly reduced by short basin memory and reduced skill in the single-valued QPF.

Improving Land Data Assimilation Performance with a Water Budget Constraint

AbstractA weak constraint is introduced in ensemble Kalman filters to reduce the water budget imbalance that occurs in land data assimilation. Two versions of the weakly constrained filter, called the weakly constrained ensemble Kalman filter (WCEnKF) and the weakly constrained ensemble transform Kalman filter (WCETKF), are proposed. The strength of the weak constraint is adaptive in the sense that it depends on the statistical characteristics of the forecast ensemble. The resulting filters are applied to assimilate synthetic observations generated by the Noah land surface model over the Red Arkansas River basin. The data assimilation experiments demonstrate that, for all tested scenarios, the constrained filters produce analyses with nearly the same accuracy as unconstrained filters, but with much smaller water balance residuals than unconstrained filters.

Soil Moisture Sensitivity to NRL-Blend High-Resolution Precipitation Products: Analysis of Simulations With Two Land Surface Models

<para xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> We examine the Naval Research Laboratory (NRL) blended satellite (NRL-Blend) High-Resolution Precipitation Product (HRPP) as a proxy for a Global Precipitation Mission (GPM)-era HRPP by using the NRL-Blend for precipitation forcing in land surface models (LSM). We use the existing (late 2008) constellation of low Earth orbiting (LEO) microwave-based satellite platforms as a baseline to examine the impact of omitting several satellite and sensor types from future GPM-era HRPPs. A response of 1-m soil water content (SWC) to different precipitation forcing represented by six NRL-Blend satellite/sensor omission scenarios was investigated using simulations over the central United States with the Noah and Mosaic land surface models (LSM). The LSMs were integrated over a domain encompassing the Arkansas-Red River basin, using the North American Land Data Assimilation System (NLDAS) atmospheric forcing (except for precipitation). Both spatial and temporal statistical properties of the SWC response were examined. Both LSMs predicted a rather consistent geographical response of the 1-m SWC to different precipitation inputs, having positive/negative SWC monthly mean anomalies in western/eastern parts of the domain. The biggest impact was due to the omission of either the crosstrack microwave sounders, or the morning local time crossing satellites. On the other hand, omission of afternoon local time crossing satellites in the NRL-Blend resulted in the smallest impact upon the soil moisture simulated both with the Noah and Mosaic models. Although the relative magnitude of these SWC changes is small, these results suggest the importance of the crosstrack microwave sounders for future GPM constellations. </para>

A Multiscale Ensemble Filtering System for Hydrologic Data Assimilation. Part II: Application to Land Surface Modeling with Satellite Rainfall Forcing

Abstract Part I of this series of studies developed procedures to implement the multiscale filtering algorithm for land surface hydrology and performed assimilation experiments with rainfall ensembles from a climate model. However, a most important application of the multiscale technique is to assimilate satellite-based remote sensing observations into a land surface model—and this has not been realized. This paper focuses on enabling the multiscale assimilation system to use remotely sensed precipitation data. The major challenge is the generation of a rainfall ensemble given one satellite rainfall map. An acceptable rainfall ensemble must contain a proper multiscale spatial correlation structure, and each ensemble member presents a realistic rainfall process in both space and time. A pattern-based sampling approach is proposed, in which random samples are drawn from a historical rainfall database according to the pattern of the satellite rainfall and then a cumulative distribution function matching procedure is applied to ensure the proper statistics for the pixel-level rainfall intensity. The assimilation system is applied using Tropical Rainfall Measuring Mission real-time satellite rainfall over the Red–Arkansas River basin. Results show that the ensembles so generated satisfy the requirements for spatial correlation and realism and the multiscale assimilation works reasonably well. A number of limitations also exist in applying this generation method, mainly stemming from the high dimensionality of the problem and the lack of historical records.

Characteristics of warm season precipitating storms in the Arkansas–Red River basin

Analysis of a multisensor precipitation product enables us to extract the precipitation from individual storms in the Arkansas–Red River drainage basin over a period of 11 years. We examine the year‐to‐year and intraseasonal variations of storm numbers, duration, sizes, and precipitation in the data set. Intraseasonal variations in numbers of storms exceed their year‐to‐year variations. More mountainous regions had greater numbers of storms than flatter regions. Most storms are small, last less than 2 h, and produce modest amounts of precipitation. The maximum size of storms and the number of storms are negatively correlated on a yearly basis. Midsummer months had a greater percentage of smaller storms but the storms were of longer average duration. We can roughly divide the storms into three different types, single ordinary cell storms, multiple storms (includes supercells), and mesoscale convective systems, and look at their year to year and intraseasonal variability in the data set. The most storms occur around 1700 local time but the most precipitation falls around 0100 local time. Storm duration was the most important factor determining how much precipitation storms generate per cell. We do not find that drought years or years with abundant precipitation had any particular characteristics but occur as a result of simultaneous occurrence of several features.

A Multiscale Ensemble Filtering System for Hydrologic Data Assimilation. Part I: Implementation and Synthetic Experiment

Abstract The multiscale autoregressive (MAR) framework was introduced in the last decade to process signals that exhibit multiscale features. It provides the method for identifying the multiscale structure in signals and a filtering procedure, and thus is an efficient way to solve the optimal estimation problem for many high-dimensional dynamic systems. Later, an ensemble version of this multiscale filtering procedure, the ensemble multiscale filter (EnMSF), was developed for estimation systems that rely on Monte Carlo samples, making this technique suitable for a range of applications in geosciences. Following the prototype study that introduced EnMSF, a strategy is devised here to implement the multiscale method in a hydrologic data assimilation system, which runs a land surface model. Assimilation experiments are carried out over the Arkansas–Red River basin, located in the central United States (∼645 000 km2), using the Variable Infiltration Capacity (VIC) model with a computing grid of 1062 pixels. A synthetic data assimilation experiment is performed, driven by meteorological forcing fields downscaled from the ensemble forecasts made by the NOAA/National Centers for Environmental Prediction (NCEP) Climate Forecast System (CFS). The classic full-rank ensemble Kalman filter is used as the benchmark to evaluate the multiscale filter performance, and comparisons are also made with a horizontally uncoupled filter. It is demonstrated that the multiscale filter is able to closely approximate the full-rank solution with a low computational cost (∼1/20 of the full-rank filter) in an experiment in which the top-layer soil moisture is assimilated, whereas the horizontally uncoupled filter fails to approximate the full-rank solution.

Multidecadal High-Resolution Hydrologic Modeling of the Arkansas–Red River Basin

Abstract Land surface heterogeneity and its effects on surface processes have been a concern to hydrologists and climate scientists for the past several decades. The contrast between the fine spatial scales at which heterogeneity is significant (1 km and finer) and the coarser scales at which most climate simulations with land surface models are generated (hundreds of kilometers) remains a challenge, especially when incorporating land surface and subsurface lateral fluxes of mass. In this study, long-term observational land surface forcings and derived solar radiation were used to force high-resolution land surface model simulations over the Arkansas–Red River basin in the Southern Great Plains region of the United States. The most unique aspect of these simulations is the fine space (1 km2) and time (hourly) resolutions within the model relative to the total simulation period (51 yr) and domain size (575 000 km2). Runoff simulations were validated at the subbasin scale (600–10 000 km2) and were found to be in good agreement with observed discharge from several unregulated subbasins within the system. A hydroclimatological approach was used to assess simulated annual evapotranspiration for all subbasins. Simulated evapotranspiration values at the subbasin scale agree well with predictions from a simple one-parameter empirical model developed in this study according to Budyko’s concept of “geographical zonality.” The empirical model was further extended to predict runoff and evapotranspiration sensitivity to precipitation variability, and good agreement with computed statistics was also found. Both the empirical model and simulation results demonstrate that precipitation variability was amplified in the simulated runoff. The finescale at which the study is performed allows analysis of various aspects of the hydrologic cycle in the system including general trends in precipitation, runoff, and evapotranspiration, their spatial distribution, and the relationship between precipitation anomalies and runoff and soil water storage anomalies at the subbasin scale.

Estimation of regional terrestrial water cycle using multi-sensor remote sensing observations and data assimilation

An integrated data assimilation system is implemented over the Red-Arkansas river basin to estimate the regional scale terrestrial water cycle driven by multiple satellite remote sensing data. These satellite products include the Tropical Rainfall Measurement Mission (TRMM), TRMM Microwave Imager (TMI), and Moderate Resolution Imaging Spectroradiometer (MODIS). Also, a number of previously developed assimilation techniques, including the ensemble Kalman filter (EnKF), the particle filter (PF), the water balance constrainer, and the copula error model, and as well as physically based models, including the Variable Infiltration Capacity (VIC), the Land Surface Microwave Emission Model (LSMEM), and the Surface Energy Balance System (SEBS), are tested in the water budget estimation experiments. This remote sensing based water budget estimation study is evaluated using ground observations driven model simulations. It is found that the land surface model driven by the bias-corrected TRMM rainfall produces reasonable water cycle states and fluxes, and the estimates are moderately improved by assimilating TMI 10.67 GHz microwave brightness temperature measurements that provides information on the surface soil moisture state, while it remains challenging to improve the results by assimilating evapotranspiration estimated from satellite-based measurements.

An observing system simulation experiment for hydros radiometer-only soil moisture products

Based on 1-km land surface model geophysical predictions within the United States Southern Great Plains (Red-Arkansas River basin), an observing system simulation experiment (OSSE) is carried out to assess the impact of land surface heterogeneity, instrument error, and parameter uncertainty on soil moisture products derived from the National Aeronautics and Space Administration Hydrosphere State (Hydros) mission. Simulated retrieved soil moisture products are created using three distinct retrieval algorithms based on the characteristics of passive microwave measurements expected from Hydros. The accuracy of retrieval products is evaluated through comparisons with benchmark soil moisture fields obtained from direct aggregation of the original simulated soil moisture fields. The analysis provides a quantitative description of how land surface heterogeneity, instrument error, and inversion parameter uncertainty impacts propagate through the measurement and retrieval process to degrade the accuracy of Hydros soil moisture products. Results demonstrate that the discrete set of error sources captured by the OSSE induce root mean squared errors of between 2.0% and 4.5% volumetric in soil moisture retrievals within the basin. Algorithm robustness is also evaluated for the case of artificially enhanced vegetation water content (W) values within the basin. For large W(>3 kg/spl middot/m/sup -2/), a distinct positive bias, attributable to the impact of sub- footprint-scale landcover heterogeneity, is identified in soil moisture retrievals. Prospects for the removal of this bias via a correction strategy for inland water and/or the implementation of an alternative aggregation strategy for surface vegetation and roughness parameters are discussed.

Multiple-Timescale Intercomparison of Two Radar Products and Rain Gauge Observations over the Arkansas–Red River Basin

A detailed intercomparison was performed for the period January 1998–June 1999 of three different sets of rainfall observations over the watershed covered by the National Weather Service Arkansas–Red Basin River Forecast Center (ABRFC). The rainfall datasets were 1) hourly 4-km-resolution ABRFC-produced P1 estimates, 2) 15-min 2-km resolution NOWrad estimates produced and marketed by Weather Services International Corporation (WSI), and 3) conventional hourly rain gauge observations available from the operational observing network. Precipitation estimates from the three products were compared at monthly, daily, and hourly timescales for the Arkansas–Red River basin and the Illinois River basin. Results indicate that the P1 products had a higher correlation and smaller bias relative to rain gauges than did the WSI products. The fact that the P1 estimates are bias corrected using gauges themselves makes an independent assessment difficult. WSI monthly accumulations seemed to overestimate (underestimate) total rainfall relative to gauges during the warm (cold) season. WSI and P1 estimates had very good agreement overall with correlation coefficients of daily accumulations generally greater than 0.7. The P1 hourly estimates were characterized by a large proportion of extremely light rainfall rates (less than 2 mm h−1). This is likely due to the P1 bias correction algorithm's use of sparse gauge data during low-level stratiform precipitation events. Finally, analyses of mean areal precipitation, fractional coverage, and storm total rainfall for the Illinois River basin demonstrate the potential impact of these rainfall products on hydrologic models that use these precipitation estimates as meteorological forcing.

Disaggregation of spatial rainfall fields for hydrological modelling

Abstract. Meteorological models generate fields of precipitation and other climatological variables as spatial averages at the scale of the grid used for numerical solution. The grid-scale can be large, particularly for GCMs, and disaggregation is required, for example to generate appropriate spatial-temporal properties of rainfall for coupling with surface-boundary conditions or more general hydrological applications. A method is presented here which considers the generation of the wet areas and the simulation of rainfall intensities separately. For the first task, a nearest-neighbour Markov scheme, based upon a Bayesian technique used in image processing, is implemented so as to preserve the structural features of the observed rainfall. Essentially, the large-scale field and the previously disaggregated field are used as evidence in an iterative procedure which aims at selecting a realisation according to the joint posterior probability distribution. In the second task the morphological characteristics of the field of rainfall intensities are reproduced through a random sampling of intensities according to a beta distribution and their allocation to pixels chosen so that the higher intensities are more likely to be further from the dry areas. The components of the scheme are assessed for Arkansas-Red River basin radar rainfall (hourly averages) by disaggregating from 40 km x 40 km to 8 km x 8 km. The wet/dry scheme provides a good reproduction both of the number of correctly classified pixels and the coverage, while the intensitiy scheme generates fields with an adequate variance within the grid-squares, so that this scheme provides the hydrologist with a useful tool for the downscaling of meteorological model outputs. Keywords: Rainfall, disaggregation, General Circulation Model, Bayesian analysis