Building the future by doing more together

EROSFIRE - A model-based, decision-support tool for soil erosion hazard assessment following forest wildfires
Coordinator - Jan Jacob Keizer
Programme - POCI
Execution dates - 2005-06-01 - 2008-05-31 (36 Months)
Funding Entity - FCT
Funding for CESAM - 70955 €
Total Funding - 87301 €
Proponent Institution - Universidade de Aveiro
Participating Institutions
Escola Superior Agrária de Coimbra (ESAC/IPC)

From a conceptual point of view as well as from that of explicitly testing of current ideas on hydrological and soil erosion processes, there have been major advances in quantitative soil erosion prediction from the empirical models, like the Universal Soil Loss Equation (USLE), to physically-based models like LISEM that are spatially explicit and fully integrated in a GIS-environment. Even so, soil erosion modeling continues to be surrounded by considerable uncertainty and the influence of calibration on model results remains rather pronounced. The uncertainty in model results is in part due to the inherently non-linear nature of overland flow processes as well as processes of soil detachment and transport. On the other hand, also conceptual and technical difficulties in measuring specific variables can play a role, by causing measurement errors in values used for model parameter estimation as well as in values used in model assessment. In spite of the aforesaid, a particularly constraint in soil erosion modeling is often the lack of data and information, whether as input for model parametrisation, model calibration or for model assessment. Extensive data sets are also of particular importance for model assessment, since they greatly facilitate assessing model robustness and, thereby, improving predictive confidence intervals. <br>Whilst data availability is thus an important bottleneck in soil erosion modeling, the standard practice of field erosion studies aims at measurements over rather long observation periods but at rather few locations. Besides costly and time-consuming, this practice is especially little appropriate for modeling purposes in rapidly changing environments, like following wildfires in forest areas, since subsequent rainfall events used for calibration will easily lack the required comparability. <br>In the current proposal, rainfall simulation experiments (rse) are presented as a possible way-around the referred data constraints and a comparatively cheap and fast method to gather significant amounts of data under distinct circumstances. The additional monitoring of small-scale and slope-scale plots is intended to ensure that the rse's are indeed sufficiently representative of real-world situations. The rse data are thought to provide a sufficient basis for erosion modeling at the small-plot scale and, through upscaling, for predicting erosion rates at the slope scale, with repeated rse campaigns intended to capture the possibly pronounced temporal variability in hydrological and erosion processes following wildfires. Various existing erosion models will then be compared, and the best model or models, eventually after some reformulations, is intended to provide the output that will allow erosion hazard assessment and mapping at the scale of individual hill slopes for different post-fire management scenarios. This tool is hoped to contribute to some future integrated decision-support system for post-fire forest management

Members on this project
Celeste Coelho
Jan Jacob Keizer
Principal Researcher


CESAM Funding: