Open AccessSensitivity of Latent Heat Fluxes to Initial Values and Parameters of a Land‐Surface Model
Author(s) -
Schwinger Jörg,
Kollet Stefan J.,
Hoppe Charlotte M.,
Elbern Hendrik
Publication year - 2010
Publication title -
vadose zone journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.036
H-Index - 81
ISSN - 1539-1663
DOI - 10.2136/vzj2009.0190
Subject(s) - water content , environmental science , soil science , tangent , data assimilation , soil texture , latent heat , saturation (graph theory) , vegetation (pathology) , sensitivity (control systems) , leaf area index , forcing (mathematics) , moisture , soil water , atmospheric sciences , mathematics , meteorology , geography , geology , geotechnical engineering , geometry , medicine , ecology , pathology , combinatorics , electronic engineering , biology , engineering
A tangent linear version of the complex soil–vegetation–atmosphere transfer model Community Land Model has been developed and its ability to reproduce relevant sensitivities of the modeled soil moisture and latent heat flux (LE) evolution to a number of different parameters on a short time scale (3 d) was analyzed. To this end, a series of idealized experiments for different soil moisture states and three different soil types was conducted. We found that the tangent linear model performs well for a large range of conditions. Situations in which the linear approximation potentially fails were connected with the occurrence of saturation and with the highly nonlinear parameterization of water availability to plants. The sensitivity of LE with respect to the model's initial soil moisture state was calculated and compared with the LE sensitivity to soil texture, leaf area index (LAI), and vegetation roughness length. These sensitivities were found to be highly variable with soil type and soil moisture. Our results confirm that soil texture and LAI are key parameters that have a dominant influence on modeled LE under specific environmental conditions. As a preparatory study to soil data assimilation developments, the results also serve to quantify model uncertainties due to limited knowledge of forcing processes and model parameters.