Abstract:Multivariate statistical analysis was used to explore relationships between catchment topography and spatial variability in snow accumulation and melt processes in a small headwater catchment in the Spanish Pyrenees. Manual surveys of snow depth and density provided information on the spatial distribution of snow water equivalent (SWE) and its depletion over the course of the 1997 and 1998 melt seasons. A number of indices expressing the topographic control on snow processes were extracted from a detailed digital elevation model of the catchment. Bivariate screening was used to assess the relative importance of these topographic indices in controlling snow accumulation at the start of the melt season, average melt rates and the timing of snow disappearance. This suggested that topographic controls on the redistribution of snow by wind are the most important influence on snow distribution at the start of the melt season. Furthermore, it appeared that spatial patterns of snow disappearance were largely determined by the distribution of snow water equivalent (SWE) at the start of the melt season, rather than by spatial variability in melt rates during the melt season. Binary regression tree models relating snow depth and disappearance date to terrain indices were then constructed. These explained 70-80% of the variance in the observed data. As well as providing insights into the influence of topography on snow processes, it is suggested that the techniques presented herein could be used in the parameterization of distributed snowmelt models, or in the design of efficient stratified snow surveys.

No abstract

Abstract:A sensitivity study of the subsurface flow component of the physically based distributed modelling system SHETRAN using data from a small Mediterranean mountain catchment is presented. The parameter space sampled was based on ranges of parameter values measured in field experiments. Model results were validated by comparison against outlet discharge, soil moisture reserve and phreatic surface level, using a number of criteria. The objectives of this exercise were: to explore variability in simulation response produced by uncertainty in parameter values; to use internal data to examine process representation within the model; and to attempt to reduce uncertainty in parameter estimates through use of a number of response variables and a number of criteria in model evaluation.A wide range of responses were produced from across the parameter space sampled. Furthermore, the sensitivity analysis demonstrated that interactions between parameters are complex, and that the sensitivity of individual parameters changed according to the values taken by other parameters and to the state of the system.The multi-response, multi-criteria evaluation presented herein exposed the presence of compensating errors within the model, which conventional procedures, which only consider outlet discharge, might fail to detect. This showed that parameter optima occurred in different parts of the parameter space for each of the response variables considered, although it did show that there was a region of the parameter space in which acceptable results were found for all responses.It is suggested that the fact that parameter optima were found in different parts of the parameter space for different responses indicates incorrect estimation of those model parameters not considered as part of the sensitivity study or omissions in the process representation embodied within the current SHETRAN subsurface flow model.

An evaluation of the performance of a physically-based distributed model of a small Mediterranean mountain catchment is presented. This was carried out using hydrological response data, including measurements of runoff, soil moisture, phreatic surface level and actual evapotranspiration. A-priori model parameterisation was based as far as possible on property data measured in the catchment. Limited model calibration was required to identify an appropriate value for terms controlling water loss to a deeper regional aquifer. The model provided good results for an initial calibration period, when judged in terms of catchment discharge. However, model performance for runoff declined substantially when evaluated against a consecutive, rather drier, period of data. Evaluation against other catchment responses allowed identification of the problems responsible for the observed lack of model robustness in flow simulation. In particular, it was shown that an incorrect parameterisation of the soil water model was preventing adequate representation of drainage from soils during hydrograph recessions. This excess moisture was then being removed via an overestimation of evapotranspiration. It also appeared that the model underestimated canopy interception. The results presented here suggest that model evaluation against catchment scale variables summarising its water balance can be of great use in identifying problems with model parameterisation, even for distributed models. Evaluation using spatially distributed data yielded less useful information on model performance, owing to the relative sparseness of data points, and problems of mismatch of scale between the measurement and the model grid.