Klaus Eckhardt

Comparison of two different approaches of sensitivity analysis

Abstract Due to spatial variability, budget constraints or access difficulties model input parameters always are uncertain to some extent. Therefore the knowledge of sensitive input parameters is beneficial for model development and application. It can lead to a better understanding and to better estimated values and thus reduced uncertainty. In the present paper two simple approaches of sensitivity analysis are compared by the use of the physically based, continuous time hydrological model SWAT. In both approaches, one parameter is varied at a time while holding the others fixed, but the way of defining the range of variation is different. Similar results are obtained suggesting that parameter sensitivity may be determined without the results being influenced by the chosen method. Most sensitive parameters for hydrology and water quality are the physical soil properties such as bulk density, available water capacity or hydraulic conductivity. Plant specific parameters like maximum stomatal conductance or maximum leaf area index as well as slope length, slope steepness, and curve number also show a high sensitivity. Both approaches can be considered as equivalent, as they provide the same overall ranking into more and less sensitive parameters. An identification of the sensitive parameters is possible independently from the chosen variation range.

Automatic calibration of a distributed catchment model

Parameters of hydrologic models often are not exactly known and therefore have to be determined by calibration. A manual calibration depends on the subjective assessment of the modeler and can be very time-consuming though. Methods of automatic calibration can improve these shortcomings. Yet, the high number of parameters in distributed models makes special demands on the optimization. In this paper a strategy of imposing constraints on the parameters to limit the number of independently calibrated values is outlined. Subsequently, an automatic calibration of the version SWAT-G of the model SWAT (Soil and Water Assessment Tool) with a stochastic global optimization algorithm, the Shuffled Complex Evolution algorithm, is presented for a mesoscale catchment.

Hydrologic Response to land use changes on the catchment scale

Abstract Regional land use changes due to European market policy have far reaching consequences for various landscape functions. Among others land use change influences the local water balance. Simulation models are mostly used to analyse the effect of management practices on water quality but they can also be a useful tool to quantify the hydrologic response of a catchment to different land use options. In this study the physically based, continuous time step model SWAT mod has been applied within the joint research project SFB 299 at the Giessen University to support the development of sustainable land use concepts. In a first step the model has been calibrated and validated for four mesoscale watersheds with differing land use distributions. Then the model performance for changing land use has been tested in an artificial watershed with a single crop at one time and one underlying soil type to eliminate the complex interactions of natural watersheds. In relation to forest barley produced the strongest response of the water balance components. Finally a case study for the Dietzholze watershed with two land use scenarios derived with the ProLand model has been carried out. The impact of land use change on the annual water balance was relatively small due to compensating effects in a complex catchment. The decrease of forest due to a grassland bonus amplifies the peak flow rate and thus increases the risk of flooding.

Plant parameter values for models in temperate climates

Abstract Ecological, and especially hydrological models used to assess the effects of land cover changes require various input parameters for plants. Regional model applications rely on detailed information about the properties of the vegetation, especially if process-based approaches are chosen. As raising acceptable data is a time consuming issue, scientists often use globally approximated plant parameter ranges, rather than considering published data sets. The plant parameters summarised in this overview, i.e. albedo, interception capacity, maximum leaf area index, rooting depth, plant height and stomatal conductance, can be used as data for a wide range of published ecological and hydrological models. We concentrate on a presentation of values for temperate regions in order to list a manageable amount of data. Information on plant species is grouped into four main land cover types, crops, pasture (herbs, forbs, grasses), coniferous and deciduous trees. Overall, more than 1300 values for the described parameters have been gathered and present a solid data base for future applications. Further properties of species and sites, such as stand age, basal area, stock density, plant height, mean annual precipitation, mean annual temperature, coordinates and country are given, if available. In many cases of model applications scientists used parameter spans, with no further information or testing of the distribution of data. Twenty-two of the total of 26 data sets subsumed in this data base contained sufficient values to perform a Kolmogorov–Smirnov-test. Twenty of these 22 data sets are normally distributed. In order to investigate spatial differences, the data for stomatal conductance, leaf area index and interception capacity were grouped into North American and European land cover species. Significant differences could only be determined for the leaf area index of deciduous trees and pasture species between the continents.

Parameter uncertainty and the significance of simulated land use change effects

Abstract Uncertainty in parameters characterising different land covers leads to uncertainty in model predictions of land use change effects. In this study, a new approach is presented which allows a model to be assessed to see whether it is suitable for investigating land use change scenarios in the sense that different land covers can be significantly distinguished in their effects on model output. It consists of the following steps: (a) The uncertainty in land cover-dependent parameters is quantified. (b) The model of an artificial catchment with representative characteristics and uniform land cover is established. (c) Using this artificial catchment, Monte Carlo simulations are carried out to determine the uncertainty in the model response to different land covers. (d) By comparing the results for two covers, respectively, a dimensionless test statistic, the distinction level, is calculated. The distinction level is a normalised probability that two independent realisations of land covers which are parameterised within their range of natural uncertainty will yield distinct model responses. If the distinction level is greater than or equal to 90%, the land covers are assumed to have a significantly different effect on the model output. An example of the application of the new method is provided using the eco-hydrologic model SWAT-G. The land covers forest, pasture and arable land can be significantly distinguished by their long-term means of surface runoff, groundwater recharge and streamflow. The minimum proportion of the catchment area on which land cover must change in order to obtain significantly distinct model responses depends on the land covers involved and the considered hydrologic variable. In the case of a change between pasture and forest and with regard to average streamflow, this minimum proportion amounts to about 25%, a value that compares well with the results of paired catchment studies.

SWAT-G, a version of SWAT99.2 modified for application to low mountain range catchments

Abstract The Soil and Water Assessment Tool (SWAT) is a well established distributed eco-hydrologic model. However, using the example of a mesoscale catchment in Germany it is shown that the version SWAT99.2 is not able to correctly reproduce the runoff generation in a low mountain region. The calculated contribution of the baseflow to the streamflow is far too high whereas the interflow is strongly underestimated. Alternatively, the modified version SWAT-G can be used which, as is demonstrated in this paper, yields far better results for catchments with predominantly steep slopes and shallow soils over hard rock aquifers. In the example, calibrating the model over three hydrologic years of daily streamflow, the model efficiency increases from −0.17 to +0.76. The modifications in SWAT-G allow hydrological processes to be modelled in low mountain ranges while not restraining the applicability of the model to catchments with other characteristics.