4.2 Projected soil erosion rates
Overall, the projected changes in SE rates with a BAU scenario under the future climatic conditions provided an average yearly SE rate for the next 20 years in the order of 2.9 and 3.0 t/ha/year for RCPs 4.5 and 8.5, respectively, both slightly lower than the HIST (1980-2000) average of 3.3 t/ha/year. The results agree with the projected climate indices for the area, which depict a future scenario of lowering precipitation and moderate extremes’ increment. HIST and future values of SE were comparable with the SE classification proposed by Panagos et al. (2015a) for the Alentejo region, which reported SE rates from 2.0 to 5.0 t/ha/year; but they are slightly lower than the estimated average SE rate of 1.2 t/ha/year reported by Cerdan et al. (2010) for the whole area of Portugal. In term of soil stability and sustainability, Verheijen et al.(2009) indicated a tolerable range of SE between 0.4 and 1.4 t/ha/year, recommended to maintain a sustainable equilibrium between soil formation and soil loss for the European countries. Both HIST and future SE rates estimated in our sites are higher than this proposed tolerable range but they are in agreement with the results reported by Panagos et al. (2021) which forecasted an average SE for 2050 equal to 3.7 t/ha/year in agricultural areas across Europe. Moreover, the concomitant occurrence of dry and wet extremes due to CC might exacerbate the susceptibility to desertification of the entire region (Mirzabaev et al., 2019), increasing SE, decreasing crop yields and livestock productivity, with cascading impacts on food security and nutrition (EEA, 2019). Such SE rates represent a serious environmental issue in particular for the driest areas of Europe, like the Alentejo, under agricultural land use. Agricultural areas used for cropping or animal breeding are generally characterized by a significant removal of the C fixed via primary productivity and hence have less C available to feed SOC formation processes (Soussana et al., 2010; Whitehead, 2020), respect to natural environment. In drier conditions, primary productivity is further limited by water availability, further reducing the possibility for significant C input to the soil (ref). In these conditions SE rates greater than formation rates entail an irreversible loss of fertile topsoil, specifically organic carbon and nutrients, with a consequent decrease of soil-related ecosystem services (Steinhoff-Knopp et al., 2021), impacts on the global carbon cycle and increased GHGs emissions at watershed scale, especially in managed landscapes (Lal, 2019; Chappell et al., 2016).
In this respect, the SE results obtained in the simulation for the whole basin showed an improving SE trend, in particular under the most probable scenario RCP 4.5 (Table 2), with values that remained below 2 t/ha/year for FRSE and OLIV, which together cover 33% of the studied basin. However, the SE data reported in Table 2 are the average of land units characterized by the same land cover but having different parameters relevant for SE rate magnitude, like slope, soil type, SOC and other soil physical characteristics (Dissanayake et al., 2019). However, to improve land management it is more effective to represent the results in terms of SE susceptibility maps, rather than representing them as a simple average of the analysed SE for the whole basin. The SE susceptibility maps produced for the basin could help to provide the landowners of the four LIFE sites, not only with a clear picture of which areas would be more prone to SE, thus requiring higher attention, but also which areas might be more susceptible to further increment of SE risk in the future, thus needing more stable and long term management solutions. These maps showed a comparable pattern of the susceptibility classes for the HIST and the future scenarios (RCP 4.5 and 8.5) (Fig. 5) and clearly identified those areas which might be of major concern for land managers, but they were also used to evidence in which areas CC produced a substantial variation of SE with a BAU scenario (Fig. 6). The latter shows that under both future scenarios, the yearly SE average will be lower compared to the HIST scenario, with few exceptions mainly located in the central area of the basin. Considering the RCP 4.5 scenario, 95% of the most endangered areas are characterized by Leptosols, slope >10%, and a predominant land cover of managed ecosystems (“montado” WPAS 50%, OAT 25%, OLIV 2%) and only a quarter covered by natural ecosystems (FRSE 23%), which makes these areas particularly in need to a dedicated land management plan to reduce SE effects. Moreover, comparing the maps of the SE difference between the HYST and the two future scenarios (Fig. 6), a very different average SE pattern is evident, which might be very relevant for management purposes.
According to this analysis, the magnitude of the SE rate in the studied basin changes significantly with factors such as soil erodibility, land cover, land management, and slope (Nùnes et al., 2011). Furthermore, in our study the SE susceptibility resulted more strongly influenced by soil and land use rather than morphology. In fact, the concomitant presence of WPAS on Leptosols determined the highest SE susceptibility of the whole central southern part of the basin. On one hand, Leptosols peculiarities (i.e., sandy texture, low soil’s depth, and low SOC; Ebelhar et al., 2008), negatively impact soil erodibility determining low aggregates size and stability and slow plant growth. On the other hand, the presence of the montado system (WAPS) characterized by the exploitation of multiple resources (i.e., livestock, forestry, and crops) causes severe impacts on soil, increasing SE risk despite the presence of forestry and olive plantations (Shakesby et al., 2002). Indeed, continuous livestock trampling leads to soil compaction which reduces water infiltration with a consequent increase of surface runoff (Choelo et al., 2004), while intensive farming leads to loss of soil organic matter intensifying the degradation of such yet fragile soils (Nùnes et al., 2008). Continuous livestock grazing also reduces the vegetation cover that in turn negatively affects surface runoff under different rainfall intensity events, promoting soil loss. Schnabel et al. (2009) showed that for the grazed “dehesa” systems of Extremadura, an environmental system similar to “montando”, a ground cover of at least 60% was necessary to protect the soil during exceptionally high intensity storms (I-30>40 mm·h-1). Whilst, a ground cover lower than 20% represented a threat, because soil loss occurred even in moderately intense storms. Mean SE rates related to sheet wash events was estimated to vary between 0.12 and 1.34 t/ha/year for 60% and 20% of ground cover, respectively. These estimates, corroborated by the studies of Kosmas et al. (1997) and Ceballos et al. (2002), clearly highlighted the key role of vegetation in land surface processes, which is why vegetation can be considered anecosystem services provider (Guerra et al., 2016).
Finally, the low SE susceptibility across the north eastern border of the watershed match the OLIV (medium to low SE) and FRSE (low SE) distribution, on soil formations other than Leptosols. Accordingly, forests (FRSE) proved to exert a protective action on soil (Borrelli et al., 2017) exhibiting the lowest SE rates due to their capacity to intercept the rainfall by the tree canopies, independently of the soil type and morphology that characterize the area.