Data collection calibration of PESERA against reservoir data
Whereas a lot of attention in DESIRE is given to the effects of protection and restoration measures, land use and management practices on soil erosion at the scale of farmers' plots or hill slopes, several studies have indicated that the extrapolation of soil erosion rates at the plot scale to sediment export rates at catchment scale is not straightforward (e.g. Walling, 1983; Poesen & Hooke, 1997; de Vente & Poesen, 2005; de Vente et al., 2007). Nevertheless, insight in the total sediment export at catchment scale is needed to evaluate the effects of mitigation strategies and management practices at larger spatial units but also to address off site consequences of runoff and soil erosion such as reservoir sedimentation and flooding.
To help solving this scaling problem, K.U. Leuven (partner 2) developed a database with sediment export rates from river catchments in Europe, the Mediterranean World and the regions of the DESIRE hotspot areas outside Europe. The general objective of this sediment yield (SY) database in WB 5 is to allow the calibration and validation of the (adapted) PESERA model and provide a framework to evaluate mitigation strategies at the catchment scale, considering their effects on the total sediment export.
The SY database was constructed, based on a database on published reservoir siltation rates (Verstraeten et al., 2006), data from publications, reports, PhD. and MSc. thesis's, and data from other DESIRE partners. Although the SY database continues to expand, a first version of the database contains sediment export data, measured at 1630 different locations in Europe, representing at least 26 202 catchment years of measured data. 506 of the sediment export data were derived from reservoir surveys (R), while 1124 of the sediment export data were measured at gauging stations (GS). The database covers catchment areas ranging from 0.01 km² to > 100 000 km². A detailed description of the dataset will be provided in Vanmaercke et al. (in prep.).
The compilation of the SY database is directly related to the compilation of a second database containing soil loss data from runoff plots. In this database, plot-scale runoff and soil loss data are collected through a detailed literature review of journal papers, books, PhD theses, internal project reports and through correspondence with researchers collecting runoff and soil loss data from plots. This database will also be used for model calibration and validation. However, a first analysis of the available sediment export data, confronted with the soil loss data on the plot scales, revealed some noticeable trends.
The European SY data was classified in different climatic zones, based on the LANMAP 2 classification (Mücher et al., 2006; Metzger et al., 2005). An analysis of the cumulative distribution of area-specific sediment yields (SSY, ton/km²/yr) revealed that sediment yields in the Mediterranean climatic zone are significantly (a factor 2 to 10) higher then in other climatic zones.
This draws attention to the Mediterranean region as a sediment yield hotspot. Several authors have already indicated that sediment fluxes in semi-arid regions, and more specific in the Mediterranean basin, are generally higher and more sensitive to (human) disturbances (e.g. Walling & Kleo, 1979; Woodward, 1995). The compiled database offers, however, a first way to quantify and analyse differences in a detailed way, based on a sufficient amount of measured data.
A comparison of the measured sediment export data with the soil loss rates at plot scale for the different climatic region further indicates that extrapolation of erosion rates from the plot scale to sediment export rates at the catchment scale poses difficulties, especially in the Mediterranean region. In most climatic zones, soil loss rates at the plot scale are generally higher then sediment yields at the catchment scale.
This agrees with the traditional expectation that sediment yields generally decrease with increasing catchment areas, as the probability that eroded sediments are deposited again increases with increasing catchment area. In this context, the sediment delivery ratio (SDR, %, the total sediment export, divided by the total gross erosion) is mostly expected to be lower then 100 % (e.g. Walling, 1983). However, for the Mediterranean region it was found that median sediment yields at the catchment scale are a factor ten higher then soil loss rates at the plot scale. Whereas the soil loss rates at the plot scale is generally lower then in many other climatic regions, sediment yields are generally higher.
Soil loss rates at the plot scale mostly take only soil erosion due to rill and interrill erosion into account. The confrontation with sediment export data clearly illustrates that often other sediment sources (e.g. gullies, bank erosion, and landslides) are most probably a more important source of soil erosion and sedimentation problems. Extrapolation from the plot scale to the catchment scale should therefore take this other sediment sources into account.
Many of the DESIRE study areas are included in the Mediterranean region, according to this used LANMAP2 classification (i.e. Guadalentin Basin, Mação, Rendina Basin, Crete, Nestos Basin and Eskisehir), while most other study sites outside Europe would be classified within a similar climatic region. These results therefore illustrate the importance of also focussing on the effects of land degradation and soil erosion at the catchment scale. Especially reservoir siltation might be an important off-site impact, since it has a direct link with water availability. Moreover, other studies have indicated that reservoirs in semi-arid regions are often more susceptible to siltation problems, due to their general higher sediment trapping efficiency (Vörösmarty et al., 2003).
The established sediment export database will allow comparison of erosion rates, predicted by the PESERA model, with actual sediment export rates. This comparison will serve as a basis indication where eventual other sediment sources are important and where additional attention needs to be given to the PESERA model.
This reservoir data collated by KU Leuven as part of WP 5.3, is being used to make comparisons with PESERA estimates. To do this, the PESERA model needs to be extended to route sediment from eroded areas to specific points downstream. There may be considerable de-coupling between hillslope and channel sediment transport, so that exceptional rates of hillslope erosion commonly produces massive valley sedimentation for many decades before the river system responds to balance the inputs. Here we assume that this process is occurring, in the short term, by associating sediment eroded from the land with a characteristic sediment transport distance related to the grain size of the source material. The travel distance from each cell is estimated from the soil type, interpreted as texture, and material is exponentially distributed downstream according to its travel distance. Additional material is assumed to accumulate on flood plains, so that reservoir sedimentation is strongly weighted towards proximate areas with high erosion rates, often associated with gullied river bank areas. The details of this method are still under discussion.
