Earth Observation Programs in support of Water Cycle Understanding and Applications
This presentation will provide an overview of water-related activities of the Group on Earth Observations (GEO), which is responsible for developing the Global Earth System of Systems (GEOSS). In particular, it will describe the research, development, and applications activities being undertaken under three GEO tasks including integrated data product development, support for water resource management and capacity building. The talk will place a special emphasis on the on-going applications activities in the Americas and opportunities for more collaboration across the Americas.
Soil Moisture Remote Sensing in the Canadian Land Data Assimilation System (CaLDAS)
For several years now, a major project has been underway at Environment Canada (EC) to improve the representation of land surface processes in operational environmental prediction systems. A foremost component of this effort is related to improving soil moisture analysis produced by the Canadian Land Data Assimilation System (CaLDAS). In its current operational state, CaLDAS assimilates observations from surface stations to specify terrestrial snow, soil moisture, and surface temperature, using simple techniques based on optimal interpolation. Our present focus is to include space-based remote sensing observations in CaLDAS with more sophisticated methods like variational data assimilation or ensemble Kalman filtering. We are currently examining and testing these approaches for the assimilation of passive L-band microwave data that will be available from the Soil Moisture and Ocean Salinity (SMOS) and Soil Moisture Active and Passive (SMAP) missions. Progress on this effort will be reported at the conference, with emphasis on the soil moisture first guess, on the integrated data assimilation framework that is becoming CaLDAS, and on specific aspects of the assimilation of SMOS and SMAP soil moisture data that require particular attention.
A comparison of soil moisture predicted from NCEP operational forecast and assimilation systems and retrieved from AMSR-E
Accurate assessment of global and regional soil moisture is critical in numerical weather and climate prediction systems because of their regulation of water and energy fluxes between the land surface and atmosphere over a variety of spatial and temporal scales. Since in-situ measurement of soil moisture is not available on global or regional scales, current coupled atmosphere-land forecast systems are relying on their land surface models to predict soil moisture, and the water and energy fluxes between the land surface and atmosphere. It is now widely acknowledged that different land surface models often spin up to rather different climatologies of soil moisture. The accuracy of the predicted soil moisture is not assessed on regular basis. Soil moisture retrieved from remote sensing platforms, e.g., the Advanced Microwave Scanning Radiometer (AMSR-E), have shown the potential to 'observe' soil moisture on the global scale. Even though its accuracy and uncertainty is yet to be determined, remotely-sensed soil moisture provides the first global picture of this crucial hydrological property and its spatial and temporal variations. In this study, soil moisture predicted from the NCEP operational global and regional forecast and assimilation systems, and that retrieved from AMSR-E will be compared. This diagnosis can provide information on how the prediction systems as well as the retrieval might be improved so that they can be used in actual global and regional land data assimilation systems.
EVALUATION OF A SOIL MOISTURE DATA ASSIMILATION SYSTEM OVER WEST AFRICA
A crucial requirement of global crop yield forecasts by the U.S. Department of Agriculture (USDA) International Production Assessment Division (IPAD) is the regional characterization of surface and sub-surface soil moisture. However, due to the spatial heterogeneity and dynamic nature of precipitation events and resulting soil moisture, accurate estimation of regional land surface-atmosphere interactions based sparse ground measurements is difficult. IPAD estimates global soil moisture using daily estimates of minimum and maximum temperature and precipitation applied to a modified Palmer two-layer soil moisture model which calculates the daily amount of soil moisture withdrawn by evapotranspiration and replenished by precipitation. We attempt to improve upon the existing system by applying an Ensemble Kalman filter (EnKF) data assimilation system to integrate surface soil moisture retrievals from the NASA Advanced Microwave Scanning Radiometer (AMSR-E) into the USDA soil moisture model. This work aims at evaluating the utility of merging satellite-retrieved soil moisture estimates with the IPAD two-layer soil moisture model used within the DBMS. We present a quantitative analysis of the assimilated soil moisture product over West Africa (9°N- 20°N; 20°W-20°E). This region contains many key agricultural areas and has a high agro- meteorological gradient from desert and semi-arid vegetation in the North, to grassland, trees and crops in the South, thus providing an ideal location for evaluating the assimilated soil moisture product over multiple land cover types and conditions. A data denial experimental approach is utilized to isolate the added utility of integrating remotely-sensed soil moisture by comparing assimilated soil moisture results obtained using (relatively) low-quality precipitation products obtained from real-time satellite imagery to baseline model runs forced with higher quality rainfall. An analysis of root-zone anomalies for each model simulation suggests that the assimilation of AMSR-E surface soil moisture retrievals can add significant value to USDA root-zone predictions derived from real-time satellite precipitation products.
The utility of MODIS products for downscaling AMSR-E soil moisture estimates
Recently, the use of microwave observations has been highlighted as a complementary tool for evaluating land surface properties. Microwave observations are less affected by clouds, water vapor and aerosol and also contain valuable soil moisture information. However, a critical limitation in these observations is the coarse spatial resolution attributed to the complex retrieval process. As a result, the best hope for providing high- resolution microwave-related products is to integrate visible, NIR/Infrared and microwave observations. In this context, we explore a downscaling approach for combining surface temperature and vegetation indices from higher spatial resolution MODIS (1km) and soil moisture from lower spatial resolution AMSR-E (25km) to obtain soil moisture at the MODIS scale (1km). The linkage is based on l integrating high resolution surface temperature and vegetation indices, through the triangle/trapezoid model, to provide an indication of relative variations in surface wetness conditions. These relative variations provide weighting parameters for downscaling of the large footprint (25km) of soil moisture to the MODIS scale (1km). Initial evaluation of the downscaled soil moisture product is undertaken at a series of Soil Moisture EXperiment (SMEX) sites conducted annually from 2002 to 2004 (Iowa, Georgia and Arizona). This concentrated validation yields insight into how well the proposed downscaling method integrates multi-scale data and the usefulness of MODIS observations in compensating for the coarse resolution of microwave observations.
Impact study of soil moisture on simulation of AMSR-E brightness temperature in NCEP data assimilation system
Soil moisture is one of most important factors not only in numerical weather and climate prediction by influencing water and energy exchange between land and atmosphere, but also in satellite data assimilation by determining brightness temperature simulation for satellite surface sensitive channels. This paper investigates impact of soil moisture on brightness temperature simulation for Advanced Microwave Scanning Radiometer (AMSR-E) low-frequency channels in the National Centers for Environmental Prediction (NCEP)'s Gridpoint Statistical Interpolation (GSI) analysis system. It is found that inaccurate soil moisture in the NCEP models results in large errors in brightness temperatures simulated through the JCSDA's Community Radiative Transfer Model (CRTM). For channels at C-band (6.925 GHz) and X-band (10.65 GHz), AMSR-E observed brightness temperatures contaminated by Radio-Frequency Interference (RFI) are corrected with the algorithm developed by Wu et al. (2009). Application of the soil moisture retrieved from AMSR-E in GSI and its potential to improve data assimilation will be discussed.
SCAT/ASCAT Soil Moisture Data: Enhancements in the TU Wien Method for Soil Moisture Retrieval From ERS and METOP Scatterometer Observations
Active microwave remote sensing observations of the scatterometers onboard the European Remote Sensing
(ERS) and METeorological OPerational (METOP) satellites have been proven to be valuable for monitoring
surface soil moisture globally using the so-called TU Wien change detection method. The METOP satellite
series carrying ASCAT (Advanced Scatteromer) instrument for the next 15 years will ensure the continuity of soil
moisture retrieval from scatterometers' data for more than 30 years considering the available ERS-1&2
Scatterometer (SCAT) observations dataset. With the aim of implementing a near real-time system for
operational soil moisture remote sensing at EUMETSAT, the Institute of Photogrammetry and Remote Sensing
at Vienna University of Technology (TU Wien) has developed an improved soil moisture retrieval algorithm to
cope with some of the limitations found in the earlier method. The new algorithm has been implemented on a
discrete global grid with 12.5 km quasi-equal grid spacing and includes a correction method to reduce
azimuthal anisotropy of backscatter signal, new techniques for calculation of the model parameters and
incorporates a comprehensive error modeling. The error analysis provides not only the quality information
about the product but also facilitates accurate determination of historically driest/wettest conditions during the
retrieval process. Enhancements made in the TU Wien retrieval algorithm result in a more uniform
performance of the model and, consequently, a spatially consistent soil moisture product with a better spatial
H23D-08 [Moved to H33A]
On The Recovery of Total Water Storage in Iran Using GRACE Observations
Iran has a dry climate with limited water resources and different parts of which suffer lack of seasonal and annual precipitation. Therefore, management of water resources is indeed, of a challenge in this region. Thanks to successful launch of GRACE mission in March 2002, which has provided us with a set of reliable data source for studying hydrological cycles in the earth fluid envelope up to now. Temporal changes in total water storage in the regional and global scales are the predominant reason of monthly and annual variations of the Earth gravity field which could be observed and detected by GRACE satellites. current hydrological models which are now being constructed using several data set (i.e. soil moisture data, in-situ measurements, raining and snowing record and so on) still seem to need more concentration and improvements. However GRACE data could be regarded as an external evidence to evaluate and modify these models; and even in some regions, as the only available information about the water storage. In this study, in-situ measurements of all existing piezometric wells in Iran have been used. The wells are not distributed homogeneously over the land and are not available every where. The data have been combined with soil moisture and snow cover information provided by GLDAS model in a monthly resolution and finally have been compared with GRACE observations in the period of time from October 2003 to September 2008. Results explicitly reveal an acceptable agreement between two sources of information and most likely it should be possible to fill the gaps in the land where in-situ measurements are not applicable.