Sediment transport using WaTEM/SEDEM

An estimate of net erosion and sediment transport using WaTEM/SEDEM in European Union

JRC in collaboration with University of Basel and Universite Cathilique de Louvain quantify the potential spatial displacement and transport of soil sediments due to water erosion at European scale. We computed long-term averages of annual soil loss and deposition rates by means of the extensively tested spatially distributed WaTEM/SEDEM model. Our findings indicate that soil loss from Europe in the riverine systems is about 15% of the estimated gross on-site erosion. The estimated sediment yield totals 0.164 ± 0.013 Pg yr-1 (which corresponds to 4.62 ± 0.37 Mg ha-1 yr-1 in the erosion area). The greatest amount of gross on-site erosion as well as soil loss to rivers occurs in the agricultural land (93.5%). By contrast, forestland and other semi-natural vegetation areas experience an overall surplus of sediments which is driven by a re-deposition of sediments eroded from agricultural land. Combining the predicted soil loss rates with the European soil organic carbon (SOC) stock, we estimate a SOC displacement by water erosion of 14.5 Tg yr-1. The SOC potentially transferred to the riverine system equals to 2.2 Tg yr-1 (~15%). Integrated sediment delivery-biogeochemical models need to answer the question on how carbon mineralisation during detachment and transport might be balanced or even off-set by carbon sequestration due to dynamic replacement and sediment burial.

Main findings:

  • WaTEM/SEDEM was applied to simulate soil loss and deposition rates at European scale.
  • Our findings indicate that soil loss in the riverine systems is about 15% of RUSLE2015 estimates. The Sediment Delivery Ration (SDR) i.e., the ratio between sediment yield (SY) and gross erosion, indicates that the sediment routed down the hillslopes to the riverine system accounts for 15.3% of the total eroded soil.
  • The estimated sediment yield in Europe totals 0.164 ± 0.013 Pg yr−1.
  • We estimate a SOC displacement by water erosion in Europe of 14.5 Tg yr−1.

WaTEM/SEDEM model

The long-term annual rates of soil loss, sediment transfer and deposition were modelled with WaTEM/SEDEM. The model has been extensively employed to estimate net fluxes of sediments across landscape, catchment- and regional-scale level. To the best of our knowledge, this study represents the first application at the continental scale. WaTEM/SEDEM is a spatially explicit sediment delivery model involving two components. In the first stage, the soil loss potential is computed according to the multi-parameter scheme of the Revised Universal Soil Loss Equation (RUSLE) (Eq. 1).

SL=R·K·LS2D·C·P (Eq. 1)
where SL is the mean soil loss (Mg ha-1 yr-1) which is the product of the rainfall intensity factor R (MJ mm ha-1 h-1 yr-1), the soil erodibility factor K (Mg h MJ-1 mm-1), the two-dimensional slope and slope-length factor LS2D (Desmet and Govers, 1996), the cover-management factor C (dimensionless) and the conservation support practice factor P (dimensionless).

In the second step, the displaced soil amount (gross erosion) is routed downslope across each pixel from hillslopes to the riverine systems according to the transport capacity (TC in Mg yr-1) (Eq. 2), computed on the base of topography and land cover.

TC =ktc·EPR =  ktc·R·K·(LS2D-4.1·SIR )       (Eq. 2)
where TC is the transport capacity (Mg ha-1 yr-1), ktc (m) is the transport capacity coefficient, R, K, LS2D are the aforementioned RUSLE input factors and SIR (m m-1) (Eq. 3) is the inter‐rill slope gradient computed based on Govers and Poesen (1988) (Eq. 3):

SIR = 6.8·Sg0.8       (Eq. 3)    where Sg represent the slope gradient (m m-1).

To run WaTEM/SEDEM we employed the RUSLE parameters (R-, K-, C-, P-factor) recently developed by the Joint Research Centre in collaboration with several European scientists (Panagos et al., 2015). Since topography plays a central role in the model, a high-resolution (25 m) digital elevation model (DEM) was employed. The RUSLE parameters were resampled to 25 m through a nearest neighbor resampling algorithm to obtain a set of gridded layers spatially consistent.

To optimize the WaTEM/SEDEM simulations across the large modelling area, the calibration of the ktc coefficients, reflecting the vegetation component in the transport capacity, was conducted considering large ranges of values (ktc_low range 0-0.5, in steps of 0.05; ktc_high range 20-600, in steps of 20). In addition, a range of different thresholds to define the upslope contributing area (Ac) was used (50, 100, 150 and 250 hectares). For the calibration of the model, a set of 24 catchments well distributed across Europe were employed. The catchment areas range from 2.5 to 245 km2. For each catchment ~1,300 model runs were performed to simulate the sediment yield for each possible combination of of ktc_min, ktc_high and Ac.

Following the Generalized Likelihood Uncertainty Estimation (GLUE) the optimal set of parameters for the median confidence level (ME = 0.89) is ktc_max of 20 m,  ktc_min = 10 m and an upslope contributing area (Ac) of 150 hectares.

Further considerations - Future research

Although our modelling approach presents an important step forward by allowing high resolution large-scale prediction of soil loss (25 × 25 m), supported by good calibration results, the insights gained by the analysis of the results highlight the need to further improve the calibration scheme of the model transport parameter in order to better reconcile the good agreement between predicted and measured sediment yield with the spatial patterns of erosion and deposition. For WaTEM/SEDEM to serve as an effective tool for both ex-ante and ex-post policy evaluations and to increase the current understanding of erosion effects on current carbon budgets, the way forward relies on the introduction of spatially distributed calibration procedures to more effectively capture the changes in transport capacity across the different landscape features. Moreover, future research should be directed towards improving the database of sediment yield (SY) measurements. Other geomorphological processes contributing to the catchment sediment yield – for instance, gullying, tillage erosion, bank and channel erosion and re-entrainment of landslide sediments – can be active on the landscape. Therefore, for calibration/validation purposes the use SY data of catchments dominated by interrill and rill process should be preferred.

References

Borrelli, P., Van Oost, K., Meusburger, K., Alewell, C., Lugato, E., Panagos, P. 2018. A step towards a holistic assessment of soil degradation in Europe: Coupling on-site erosion with sediment transfer and carbon fluxesEnvironmental Research, 161: 291-298.

Data

The sediment transport data are available for the public user.

Fig 1. Estimated annual average soil loss and deposition rate for the European Union based on WaTEM/SEDEM. The vertical bars show the annual gross (orange) and net (red) soil losses in each country.

Go Back To

Displaying 1 - 2 of 2 |