Wolfgang Korres

Spatial Heterogeneity of Leaf Area Index (LAI) and Its Temporal Course on Arable Land: Combining Field Measurements, Remote Sensing and Simulation in a Comprehensive Data Analysis Approach (CDAA)

The ratio of leaf area to ground area (leaf area index, LAI) is an important state variable in ecosystem studies since it influences fluxes of matter and energy between the land surface and the atmosphere. As a basis for generating temporally continuous and spatially distributed datasets of LAI, the current study contributes an analysis of its spatial variability and spatial structure. Soil-vegetation-atmosphere fluxes of water, carbon and energy are nonlinearly related to LAI. Therefore, its spatial heterogeneity, i.e., the combination of spatial variability and structure, has an effect on simulations of these fluxes. To assess LAI spatial heterogeneity, we apply a Comprehensive Data Analysis Approach that combines data from remote sensing (5 m resolution) and simulation (150 m resolution) with field measurements and a detailed land use map. Test area is the arable land in the fertile loess plain of the Rur catchment on the Germany-Belgium-Netherlands border. LAI from remote sensing and simulation compares well with field measurements. Based on the simulation results, we describe characteristic crop-specific temporal patterns of LAI spatial variability. By means of these patterns, we explain the complex multimodal frequency distributions of LAI in the remote sensing data. In the test area, variability between agricultural fields is higher than within fields. Therefore, spatial resolutions less than the 5 m of the remote sensing scenes are sufficient to infer LAI spatial variability. Frequency distributions from the simulation agree better with the multimodal distributions from remote sensing than normal distributions do. The spatial structure of LAI in the test area is dominated by a short distance referring to field sizes. Longer distances that refer to soil and weather can only be derived from remote sensing data. Therefore, simulations alone are not sufficient to characterize LAI spatial structure. It can be concluded that a comprehensive picture of LAI spatial heterogeneity and its temporal course can contribute to the development of an approach to create spatially distributed and temporally continuous datasets of LAI.

Varying applicability of four different satellite-derived soil moisture products to global gridded crop model evaluation

Abstract Satellite-derived daily surface soil moisture products have been increasingly available, but their applicability to global gridded crop model (GGCM) evaluation is unclear. This study compares four different soil moisture products with the flux tower site observation at 18 cropland sites across the world where either of maize, soybean, rice and wheat is grown. These products include the first and second versions of Climate Change Initiative Soil Moisture (CCISM-1 and CCISM-2) datasets distributed by the European Space Agency and two different AMSR-E (Advanced Microwave Scanning Radiometer–Earth Observing System)-derived soil moisture datasets, separately provided by the Japan Aerospace Exploration Agency (AMSRE-J) and U.S. National Aeronautics and Space Administration (AMSRE-N). The comparison demonstrates varying reliability of these products in representing major characteristics of temporal pattern of cropland soil moisture by product and crop. Possible reasons for the varying reliability include the differences in sensors, algorithms, bands and criteria used when estimating soil moisture. Both the CCISM-1 and CCISM-2 products appear the most reliable for soybean- and wheat-growing area. However, the percentage of valid data of these products is always lower than other products due to relatively strict criteria when merging data derived from multiple sources, although the CCISM-2 product has much more data with valid retrievals than the CCISM-1 product. The reliability of the AMSRE-J product is the highest for maize- and rice-growing areas and comparable to or slightly lower than the CCISM products for soybean- and wheat-growing areas. The AMSRE-N is the least reliable in most location-crop combinations. The reliability of the products for rice-growing area is far lower than that of other upland crops likely due to the extensive use of irrigation and patch distribution of rice paddy in the area examined here. We conclude that the CCISM-1, CCISM-2 and AMSRE-J products are applicable to GGCM evaluation, while the AMSRE-N product is not. However, we encourage users to integrate these products with in situ soil moisture data especially when GGCMs simulations for rice are evaluated.