Climatologists highlight that climate change is occurring, both in terms of air temperature and precipitation patterns. According to the Intergovernmental Panel on Climate Change’s (IPCC) Fifth Assessment Report, the Mediterranean basin is expected to become warmer and dryer due to an anthropogenic increase of greenhouse gas emissions (CO2
O, and F-gases) until the end of the 21st century [1
]. Moreover, in Mediterranean regions, future warming will probably be larger than the global mean, accompanied by a considerable decrease of total rainfall amount [2
] and more frequent high-intensity rainfall events [4
]. Future climate projections will highly influence catchments’ responses to soil erosion [6
Soil erosion by water is one of the most significant forms of land degradation, as it threatens natural ecosystems, water resources, and crop productivity. The European Commission’s Soil Thematic Strategy has identified soil erosion as a key priority for the protection of soils [10
] and call for quantitative assessments of soil loss rates at the European level [11
]. Mediterranean regions are particularly vulnerable to erosion because of the highly irregular behavior of the rainfall regime, both on spatial and temporal scales [12
]; inappropriate agricultural management practices [13
]; overgrazing [14
]; and wildfires [16
]. Also, erosion rates are higher in mountainous areas than in lowlands due to the steeper relief and the higher and more intense rainfalls.
Identification of critical soil erosion-prone areas is necessary for the implementation of appropriate mitigation measures to combat erosion [18
]. Stakeholders should also quantify potential soil erosion risk under future climate conditions for the scheduling of infrastructure projects.
Actual measurements of soil erosion rates is a costly and time-consuming process that is usually limited to small experimental sites [21
].To this end, plenty of erosion prediction models were developed during the last few decades and classified on different spatial/temporal scales and various levels of complexity [26
].Erosion prediction models use mathematical equations so as to express the relationships between natural factors (e.g., rainfall, land cover, vegetation type, geological subsoil, and topography) and the erosion process. The most well-known models that have been widely used are USLE [27
], RUSLE [28
], RMMF [29
], WEPP [30
], and EPM (the Gavrilovič method) [31
The quantitative erosion prediction model of Gavrilovič was developed using field work and laboratory experiments to determine the range of values for its parameters. It was widely used in former Yugoslavia since the early 70s, and later in the Balkan region and Central Europe. It has also been recognized as the most quantitative of all the semi-quantitative models [32
]. Several studies in the mountainous catchments of Greek territory which compare the Gavrilovič prediction model results with actual measurements shows that it gives satisfactory results [33
In recent years, climate models’ projections have been used to assess soil erosion under climate change [37
]. Understanding the uncertainties of several climatological parameters is of major interest in recent climate studies [39
]. It should be noted that the ability of models to represent climate conditions should be examined prior to their use in impact assessment studies [40
Investigation of the effects of climate change on soil erosion using high-resolution Regional Climate Models (RCMs) are limited in mountainous catchments of the Mediterranean region [44
], while no such research has been carried out in Greece.
The main object of the current research is to quantify the effect of climate change on soil erosion in a mountainous catchment of Central Greece using the erosion prediction model of Gavrilovič and climate simulation of the RegCM3regional climate model.
Firstly, the ability of the RegCM3 model to simulate the climate condition (precipitation and temperature) in the study area was evaluated using criteria of the RMSE
. Results from the calculation of the RMSE
between the simulated and observed precipitations (mm) during the period of 1974–2000 are given in the figure below (Figure 3
). The model generally underestimated monthly precipitation in most cases except for January and March, where a slight overestimation can be observed. Also, it can be seen that better simulation was achieved in spring and summer.
Additionally, the above-mentioned evaluation criteria were calculated regarding temperature (°C) data (Figure 4
). It was found that the model overestimated mean monthly temperatures for almost all months, except for September and October. Moreover, as revealed from the analysis, better temperature simulation was achieved for winter and autumn.
Analysis of the land cover showed that dense forest cover was approximately 35.2% of the total catchment area, and thin forests were 24.7%, agricultural crops 18%, pastures 8.5%, dense shrubs 4.9%, scarce shrubs 3%, barren land 4.6%, and settlements 1.1%. In order to determine the coefficient of land cover (x
), based on Table 1
, the value 0.125 were was to the dense forest, 0.20 to thin forests, 0.25 to dense shrubs, 0.60 to scarce shrubs, 0.7 to pastures, 0.8 to agricultural crops, 0.9 to barren land, and 0 to settlements.
Regarding the geology of the study area, it was found that it mainly consisted of limestone (47.2%) and flysch (44.5%), followed by neogene rocks (6.5%) and alluvial deposits (1.8%). The following values of soil erodibility (y
coefficient) were assigned to each petrographic formation: 0.8 to limestone, 1.15 to flysch, 1.55 to neogene rock, and 1.0 to alluvial deposits. The coefficients of land cover (x
) and soil erodibility (y
) were selected based on literature [33
Also, following the field observations, the value 0.6 was given to the φ
coefficient. The catchment area (F coefficient) was found to be equal to 136.4 km2
after processing the DEM of the study area through GIS techniques [9
], and the slopes were found to be rather intense, ranging from 1.5% to 89.3%. The land cover, geology, and slope (%) maps are shown in the next figure (Figure 5
The climate was evaluated from the precipitation (mm) and temperature (°C) time-series data of the RegCM3 model for the periods 1974–2000 and 2074–2100. The analysis highlighted a decrease (−21.2%) in annual precipitation (mm) and a significant increase (+3.6 °C) in mean annual temperature until the end of the 21st century (Figure 6
). Specifically, the annual precipitation was expected to decrease from 1071.2 mm (1974–2000) to 874.1 mm (2074–2100) and the mean annual temperature was expected to increase from 9.3 °C to 13.0 °C. The biggest decrease in precipitation level will be in the months of spring and autumn, whereas the temperature increase will be greater in summer and winter.
Using the above-mentioned parameter and EPM formula (Gavrilovič method), the soil loss (m3
/year) and catchment erosion rate (m3
) were estimated for the baseline (1974–2000) and future (2074–2100) period. Finally, the results were compared between them so as to quantify the effects of climate change on soil erosion. It was noticed that soil loss would decrease by7920.5m3
/year (−4.9%) and, consequently, the erosion rate by 58.1 m3
. Detailed results can be seen in Table 2
According to studies which compared the results of the EPM model with actual measurements, it was found that the model overestimated soil loss by 10% [32
]. The erosion rate in the study area was considered rather high in comparison with other areas of the Greek territory [59
]. Despite the high forest cover, the vulnerable geological subsoil and the steep slopes were found to favor the development of erosion phenomena.
Nowadays, there is an urgent need for reliable future climate projections for rational catchment management and infrastructure project scheduling. The most modern tool used to simulate future climate conditions is the regional climate model (RCM). However, using models can introduce uncertainty in relation to a number of significant temporal and spatial scaling issues of the input data. Generally, RCMs underestimate monthly precipitation and overestimate temperature, whereas better simulation can be achieved in spring and summer for precipitation and winter and autumn for temperature data [60
Concerning future climate conditions in the Mediterranean region [60
], and especially Greece [3
], the climate is expected to be warmer and dryer. Additionally, the temperature in Central Europe is projected to increase between +2 °C and +5 °C. Even though annual precipitation is projected to increase up to +10%, most RCMs project a significant decrease of precipitation in summer [66
]. Moreover, extreme precipitations will be more frequent [5
], as well as torrential flooding and debris-flow phenomena.