Observation uncertainty of satellite soil moisture products determined with physically-based modeling
Wanders, N. ; Karssenberg, D. ; Bierkens, M.F.P. ; Parinussa, R. ; Jeu, R. de; Dam, J.C. van; Jong, S. de - \ 2012
Remote Sensing of Environment 127 (2012). - ISSN 0034-4257 - p. 341 - 356.
passive microwave measurements - improving runoff prediction - vegetation optical depth - ers scatterometer - amsr-e - retrieval - assimilation - validation - algorithm - index
Accurate estimates of soil moisture as initial conditions to hydrological models are expected to greatly increase the accuracy of flood and drought predictions. As in-situ soil moisture observations are scarce, satellite-based estimates are a suitable alternative. The validation of remotely sensed soil moisture products is generally hampered by the difference in spatial support of in-situ observations and satellite footprints. Unsaturated zone modeling may serve as a valuable validation tool because it could bridge the gap of different spatial supports. A stochastic, distributed unsaturated zone model (SWAP) was used in which the spatial support was matched to these of the satellite soil moisture retrievals. A comparison between point observations and the SWAP model was performed to enhance understanding of the model and to assure that the SWAP model could be used with confidence for other locations in Spain. A timeseries analysis was performed to compare surface soil moisture from the SWAP model to surface soil moisture retrievals from three different microwave sensors, including AMSR-E, SMOS and ASCAT. Results suggest that temporal dynamics are best captured by AMSR-E and ASCAT resulting in an averaged correlation coefficient of 0.68 and 0.71, respectively. SMOS shows the capability of capturing the long-term trends, however on short timescales the soil moisture signal was not captured as well as by the other sensors, resulting in an averaged correlation coefficient of 0.42. Root mean square errors for the three sensors were found to be very similar (± 0.05 m3m- 3). The satellite uncertainty is spatially correlated and distinct spatial patterns are found over Spain.
Scatterometer-Derived Soil Moisture Calibrated for Soil Texture With a One-Dimensional Water-Flow Model
Lange, R. de; Beck, R. ; Giesen, N. van de; Friesen, J. ; Wit, A.J.W. de; Wagner, W. - \ 2008
IEEE Transactions on Geoscience and Remote Sensing 46 (2008)12. - ISSN 0196-2892 - p. 4041 - 4049.
ers scatterometer - near-surface - assimilation - retrieval - validation - space
Current global satellite scatterometer-based soil moisture retrieval algorithms do not take soil characteristics into account. In this paper, the characteristic time length of the soil water index has been calibrated for ten sampling frequencies and for different soil conductivity associated with 12 soil texture classes. The calibration experiment was independently performed from satellite observations. The reference soil moisture data set was created with a I-D water-flow model and by making use of precipitation measurements. The soil water index was simulated by applying the algorithm to the modeled soil moisture of the upper few centimeters. The resulting optimized characteristic time lengths T increase with longer sampling periods. For instance, a T of 7 days was found for sandy soil when a sampling period of I day was applied, whereas an optimized T-value of 18 days was found for a sampling period of 10 days. A maximum rmse improvement of 0.5% vol. can be expected when using the calibrated T-values instead of T = 20. The soil water index and the differentiated T-values were applied to European Remote Sensing (ERS) satellite scatterometer data and were validated against in situ soil moisture measurements. The results obtained using calibrated T-values and T = 20 did not differ (r = 0.39, rmse = 5.4% vol.) and can be explained by the averaged sampling period of 4-5 days. The soil water index obtained with current operational microwave sensors [Advanced Wind Scatterometer (ASCAT) and Advanced Microwave Scanning Radiometer-Earth Observation System] and future sensors (Soil Moisture and Ocean Salinity and Soil Moisture Active Passive) should benefit from soil texture differentiation, as they can record on a daily basis either individually or synergistically using several sensors. The proposed differentiated characteristic time length enables the continuation of the soil water index of sensors with varying sampling periods (e.g., ERS-ASCAT).
Crop model data assimilation with the Ensemble Kalman filter for improving regional crop yield forecasts
Wit, A.J.W. de; Diepen, C.A. van - \ 2007
Agricultural and Forest Meteorology 146 (2007)1-2. - ISSN 0168-1923 - p. 38 - 56.
remote-sensing data - soil-moisture - spatial variability - ers scatterometer - united-states - scales - parameters - wheat - reflectances - uncertainty
Uncertainty in spatial and temporal distribution of rainfall in regional crop yield simulations comprises a major fraction of the error on crop model simulation results. In this paper we used an Ensemble Kalman filter (EnKF) to assimilate coarse resolution satellite microwave sensor derived soil moisture estimates (SWI) for correcting errors in the water balance of the world food studies (WOFOST) crop model. Crop model simulations with the EnKF for winter wheat and grain maize were carried out for Spain, France, Italy and Germany for the period 1992¿2000. The results were evaluated on the basis of regression with known crop yield statistics at national and regional level. Moreover, the EnKF filter innovations were analysed to see if any systematic trends could be found that could indicate deficiencies in the WOFOST water balance. Our results demonstrate that the assimilation of SWI has clearly improved the relationship with crop yield statistics for winter wheat for the majority of regions (66%) where a relationship could be established. For grain maize the improvement is less evident because improved relationships could only be found for 56% of the regions. We suspect that partial crop irrigation could explain the relatively poor results for grain maize, because irrigation is not included in the model. Analyses of the filter innovations revealed spatial and temporal patterns, while the distribution of normalised innovations is not Gaussian and has a non-zero mean indicating that the EnKF performs suboptimal. The non-zero mean is caused by differences in the mean value of the forecasted and observed soil moisture climatology, while the excessive spread in the distribution of normalised innovations indicates that the error covariances of forecasts and observations have been underestimated. These results clearly indicate that additional sources of error need to be included in the simulations and observations.