Criteria to Map Areas of High Risk of Soil Hydric Erosion in Portugal using USLE

The Portuguese spatial planning legislation includes legal restrictions to land use in order to preserve the ecosystems. These restrictions are framed by the legal structure called National Ecological Reserve (NER), and have associated a cartographic representation. Among the land use protection areas included in the NER are the Areas of High Risk of Soil Hydric Erosion (AHRSHE). Our goal is to improve the models and derived cartography and to use the enhanced maps as a basis to test and apply new and more advanced technologies, data and methods.

Currently, AHRSHE are determined based on USLE. The computation of the LS factor in this equation has been a challenging issue and, since this action is a legal responsibility of the municipalities, we could face a situation where different municipalities use different methodologies and, eventually, the results being not comparable. Thus, efforts have being made in order to produce a common methodology to standardise and enhance the cartographic representation of the LS, namely, by improving its accuracy and precision and by harmonizing and making it compatible with the other USLE factors. For this purpose, several methods of LS computation have been tested to evaluate soil loss risk in different geomorphic contexts. Based on the test results USLE's second revision, RUSLE2 (USDA, 2008), was selected together with imposing a maximum value to unorganised runoff length (L).

The results of using RUSLE2 might be affected by the lack of information on detailed soil properties caused by different geomorphological contexts and the lack of resolution of the Digital Terrain Model (DTM) to accurately identify the AHRSHE. The lack of DTM resolution affects the slope values (S), the shape of the hydrographic network and, above all, the delimitation of the disorganized flow domain, where AHRSHE are mapped.

In order to reach an acceptable solution, tests were made with varying maximum unorganized runoff length (L) and using different formulas to determine S, according hillslope values and rainfall regime. The test results show that the more accurate LS is obtained when L is limited to 305 m and S is calculated according to slope thresholds: below and above 9% (Panagos, et al., 2015) or above 18% (Liu, 1994; 2002), and excluding areas where the USLE is not applicable, like plane surfaces, water, or surfaces with high slopes.

Another conclusion was that small resolution DTM are inappropriate which lead us to use in the tests a 10m pixel DTM. Even so, and in order to prevent unjustified land use restrictions, we suggest the need to validate the results (by sampling), at least in specific geomorphologic contexts. Otherwise, the likelihood to get biased results, with adverse practical effects, will be high.

The shape and accuracy of AHRSHE depend on the methodologies and georeferenced data used. Thus, we intend to use, in the near future, a very-high resolution DTM derived from aerial LiDAR and to work on the identification of differentiated geomorphological contexts in each municipality in order to further improve the AHRSHE mapping, which have substantial impacts in the NER.