Search

In this study, it is shown that the complexity of Soil Vegetation Atmosphere Transfer (SVAT) models leads to an equifinality of functional behaviour - many parameterizations from many areas of the parameter space lead to very similar responses. Individual parameters derived by calibration (i.e. model inversion) against limited measurements are, therefore, highly uncertain. Due to the non-linear internal behaviour of SVAT models, aggregation of uncertainly known parameter fields to parameterize landscape scale variability in surface fluxes will yield highly uncertain predictions. A disaggregation approach suggested by Beven (1995) requires that the land surface be represented by a linear sum of a number of representative parameterizations or functional types. This study explores the nature of the parameter space in terms of a simple definition of functional behaviour. Parameter interactions producing similar predicted behaviours are investigated through application of Principal Component Analyses. These reveal the lack of a dominant global interaction indicating the presence of highly complex parameter interactions throughout the feasible parameter space.

In this study, it is shown that the complexity of Soil Vegetation Atmosphere Transfer (SVAT) models leads to an equifinality of functional behaviour - many parameterizations from many areas of the parameter space lead to very similar responses. Individual parameters derived by calibration (i.e. model inversion) against limited measurements are, therefore, highly uncertain. Due to the non-linear internal behaviour of SVAT models, aggregation of uncertainly known parameter fields to parameterize landscape scale variability in surface fluxes will yield highly uncertain predictions. A disaggregation approach suggested by Beven (1995) requires that the land surface be represented by a linear sum of a number of representative parameterizations or functional types. This study explores the nature of the parameter space in terms of a simple definition of functional behaviour. Parameter interactions producing similar predicted behaviours are investigated through application of Principal Component Analyses. These reveal the lack of a dominant global interaction indicating the presence of highly complex parameter interactions throughout the feasible parameter space.

Current ‘physically based’ soil-vegetation-atmosphere transfer (SVAT) schemes use increasingly complex descriptions of the physical mechanisms governing evapotranspiration fluxes, thereby requiring the specification of a large number of parameters controlling the vertical fluxes over a single homogeneous area. Recent attention towards the incorporation of sub-grid scale spatial variability in SVAT parameterisations promises to increase the number of parameters for these models. In this paper, it is demonstrated that a simple patch scale SVAT model still permits too many degrees of freedom in terms of fitting the model predictions to calibration or validation data; it is shown that good model fits may be achieved in many areas of the parameter space. Using a Monte Carlo framework, a sensitivity analysis is performed for simulations of data sets from FIFE and Amazonian sites. This is employed to evaluate the role of each parameter for each forcing dataset, and to identify the controlling and redundant parameters and processes. The results suggest that equifinality of parameter sets in calibration to field data must be expected, that there will be a consequent uncertainty in predictive capability and that more emphasis will be required on identifying the critical controls on evapotranspiration in extending predictions from patch to landscape scale in different environments.

Estimation of regional latent and sensible heat fluxes requires the upscaling of local models. One approach is to develop soil-vegetation-atmosphere transfer (SVAT) schemes at different scales (field, catchment and regional scale; e.g. Famiglietti and Wood (1994)). SVAT schemes at larger scales try to aggregate latent and sensible heat fluxes based on knowledge about the spatial distribution (GIS based) or the statistical-dynamic description of atmospheric forcing and internal state variables. Before one can approach this important problem of upscaling, it is necessary to study the impact of the parameters of the field scale SVAT model on the prediction of local fluxes. In this paper, we show results of a Monte Carlo sensitivity analysis performed on a local SVAT model, with data retrieved during the Autumn of 1995 in an experimental catchment in Belgium. During this field campaign, several parameters of the energy and water balance were measured by means of a Bowen ratio system, while soil moisture data were retrieved on a regular basis using an automatic TDR system. The results of the performed sensitivity analysis are presented and evaluated; conclusions are drawn with respect to the upscaling of local SVAT models to the larger scale.

Estimation of regional latent and sensible heat fluxes requires the upscaling of local models. One approach is to develop soil-vegetation-atmosphere transfer (SVAT) schemes at different scales (field, catchment and regional scale; e.g. Famiglietti and Wood (1994)). SVAT schemes at larger scales try to aggregate latent and sensible heat fluxes based on knowledge about the spatial distribution (GIS based) or the statistical-dynamic description of atmospheric forcing and internal state variables. Before one can approach this important problem of upscaling, it is necessary to study the impact of the parameters of the field scale SVAT model on the prediction of local fluxes. In this paper, we show results of a Monte Carlo sensitivity analysis performed on a local SVAT model, with data retrieved during the Autumn of 1995 in an experimental catchment in Belgium. During this field campaign, several parameters of the energy and water balance were measured by means of a Bowen ratio system, while soil moisture data were retrieved on a regular basis using an automatic TDR system. The results of the performed sensitivity analysis are presented and evaluated; conclusions are drawn with respect to the upscaling of local SVAT models to the larger scale.