- Email: [email protected]
Evaluation of water erosion at a mountain catchment in Poland using the G2 model
Catena 164 (2018) 116–124
Contents lists available at ScienceDirect
Catena journal homepage: www.elsevier.com/locate/catena
Evaluation of water erosion at a mountain catchment in Poland using the G2 model ⁎
⁎
Wiktor Haleckia, , Edyta Krukb, , Marek Ryczekb,
T
⁎
a
Department of Biometry and Forest Productivity, Faculty of Forestry, University of Agriculture in Krakow, Al. 29 Listopada 46, 31-425 Kraków, Poland Department of Land Reclamation and Environmental Development, Faculty of Environmental Engineering and Land Surveying, University of Agriculture in Krakow, 30059 Kraków, Al. Mickiewicza 24-28, Poland b
A R T I C L E I N F O
A B S T R A C T
Keywords: Erosion risk assessment G2 model Land use Poland Soil loss Surface-runoff
The Western Carpathians region in southern Poland is characterized by high erosion risk due to steep slopes, flysch formation and intense precipitation with frequent storm events. The G2 model, based on the principles of the Universal Soil Loss Equation (USLE), was used to investigate soil erosion assessment, except that the yearly rainfall erosivity factor was substituted for by the monthly one. The plant cover factor was determined based on the CORINE land cover 2012 database and field observations of vegetation stages. Slope intercept was estimated by applying a Sobel filter. Terrain properties were calculated from a 5 m DEM of the area. Modeling investigations were carried out in the agricultural basin of the mountain catchment (1.47 km2) in the years 2011–2014. Rainfall data were collected from weather stations, and soil properties were measured in 43 locations. The G2 model estimated total annual soil loss as between 3.37 Mg ha−1 (2012) and 31.05 Mg ha−1 (2014). The erosive events that contributed most to yearly erosion occurred in May (2014: 80.40% of yearly total) and June (2013: 57.08%). Redundancy analysis based on land-use types provided factors affecting soil erosion by water. In conclusion, the G2 model was useful in erosion estimation in a steep-sloped agricultural basin with a variable hydrological regime.
1. Introduction Water erosion of soil is a complex process during which fertile topsoil is detached, carried away and deposited in another location (Ballabio et al., 2017). It causes leaching of nutrients, terrain surface deformations and deterioration of water quality in catchments as well as silting of water structures and water supply and drainage systems (Żmuda et al., 2005). Even slight erosion negatively affects farming conditions and yield, impedes agrotechnical treatments, and may exclude the entire affected area from agriculture (Halecki et al., 2016). The intensity of erosion events is shaped by physiographic and hydrological factors prevailing in a specific catchment area (Verstraete and Poesen, 2001; Zabaleda et al., 2007), and by land use and plant cover (Podwojewski et al., 2008; Mao and Cherkauer, 2009). It also highly depends on geological factors, soil type, and climate (Nadal-Romero et al., 2008). The accuracy of soil loss evaluation with models depends largely on how the model parameters describe important characteristics of the catchment. Many researchers focus on the assessment of the parameters provided by the Geographic Information System (GIS) and remote
⁎
sensing data. These techniques are useful for quick evaluation of the spatial distribution of erosion within large areas, and can cover remote areas where no measurements are actually conducted (Dabral et al., 2008; Bahadur, 2009). The parameters estimated with these techniques do not differ much from those obtained from field measurements (Lee and Choi, 2010). Additionally, new geostatistical methods allow for development of maps with predicted data values that previously have had to be developed from highly time-consuming and expensive point observations (Burrough, 2001). Telemetry combined with GIS techniques and geostatistical methods enable the implementation of holistic and integrated visions of sustainable development. The most important climatic factor is precipitation, and its frequency, duration, and intensity are critical for erosion in a catchment. Water erosion processes play an important role in terrain shaping causing transformations that under natural conditions result in formation of new forms of relief. In Poland, these processes are particularly visible in mountain and foothill areas. The literature provides many generations of erosion models describing water runoff and erosion events that differ in their limiting factor(s). The most popular and widely used is the Universal Soil Loss
Corresponding authors. E-mail addresses: [email protected] (W. Halecki), [email protected] (E. Kruk), [email protected] (M. Ryczek).
https://doi.org/10.1016/j.catena.2018.01.014 Received 21 August 2017; Received in revised form 9 January 2018; Accepted 15 January 2018 0341-8162/ © 2018 Elsevier B.V. All rights reserved.
Catena 164 (2018) 116–124
W. Halecki et al.
Fig. 1. Location of the investigated catchment of the Mątny stream.
generation among all the available water erosion models of the USLE family. There has been a lack of investigation on its use in Polish conditions. Therefore, the objectives of this study were (1) to apply the G2 model to quantitative assessment of water erosion in an agricultural mountain catchment and (2) to evaluate the effect of soil and water conservation (SWC) on total soil eroded mass in a highly erosion-prone landscape.
Equation model (USLE) developed in a series of papers by Wischmeier and Smith (1978); Flanagan et al. (2003); Laflen and Moldenhauer (2003); Perović et al. (2013); and Kruk et al. (2016). The model formula, expressed as a logical product that combines all main natural and anthropogenic factors shaping the type and extent of soil erosion is as follows:
A = RKLSCP
(1)
2. Materials and methods
where: A — mean annual weight of eroded soil per unit area (Mg ha−1 y−1), R — mean annual erosivity of rainfall and runoff (MJ mm ha−1 h−1 y−1), K — soil erodibility (Mg ha h MJ−1 ha−1 mm−1), L — slope length coefficient (dimensionless), S — slope steepness coefficient (dimensionless), C — coefficient determining the type of crops and land use (dimensionless), and P — soil and water conservation (dimensionless). The USLE model is used to estimate total annual soil loss within a multi-year period. It has been constantly modified and amended as research progress is made and the availability and quality of software are improved. The following modifications are worth mentioning: MUSLE (Smith et al., 1984; Zhang et al., 2009; Cârdei, 2010, Kruk, 2017), RUSLE (Park et al., 2011; Mhangara et al., 2012; Kumar and Kushwaha, 2013), USLE-M (Kinnell and Risse, 1998; Kinnell, 2016), and USLE-MM (Bagarello et al., 2015). In 2012, G2, a new erosion model based on the USLE structure was developed at the Aristotle University of Thessaloniki and Joint Research Centre of the European Commission under the GEOLAND2 project. The model formula was elaborated by Panagos et al. (2012, 2014b, 2015a), Karydas et al. (2014, 2015), Karydas and Panagos (2016) and Zdruli et al. (2016). It is used to estimate the erosion extent on a monthly basis. Several published studies of G2 implementation, e.g. in the cross-borders basin of the Strymonas/Struma river (Greece and Bulgaria), in the basins of the Ishmi-Erzeni river and Korce region (Albania), in the Mediterranean islands of Crete (Greece) and Cyprus, have yielded realistic results (Panagos et al. 2014a; Panagos et al. 2015a,c, Karydas and Panagos, 2016, 2018, Zdruli et al. 2016). The choice of erosion modeling method usually depends on availability of proper data and the purpose of the simulation. An important advantage of the presented models is their simplicity and low requirements concerning the input data as compared with other erosion models such as SWAT — Soil and Water Assessment Tool (Ustun, 2008). The G2 model can assess soil loss in cropland and forest-dominated land uses. We hypothesized that systematically enhancing soil conservation practices might substantially reduce soil loss in consecutive years. We decided to use the G2 model because it was the newest
2.1. Study area description The study was conducted in the Outer Western Carpathians, in the southern region of Małopolska Province (Fig. 1). The terrain comprises low and medium height mountains with peak heights ranging from 617.6 m a.s.l. to 732.0 m a.s.l. The lowest point is situated 490.0 m a.s.l. The mean height of the catchment is 582.66 m a.s.l. The slope distribution is as follows: < 5%, 4.08%; 5–10%, 18.37%; 10–18%, 45.58%; 18–27%, 21.09%; and > 27%, 10.88%. The weighted mean slope for the entire catchment is 16.28%. Study area was characterized by a deep or relatively deep soil profile. The length of the main watercourse is 2.37 km and its average slope 5.7%. The river network density is 2.96 km km−2. The Mątny stream (1.47 km2) flows into the Mszanka river in the Skiby hamlet at E 20°9′2.35″, N 49°37′30.52″. The catchment land use structure is dominated by grassland (73.5%); an arable land cover of 14.3%, including spring oat (Avena sativa) — 7.3%, potatoes (Solanum tuberosum) — 4.3%, common wheat (Triticum aestivum) — 2.7%; forests account for 9.5%; and urban areas 2.7%. The catchment area is cut by a network of dirt roads. Most of them are deeply furrowed and tend to transform into water-carrying streams during and after rain events. 2.2. Climatic conditions and soil of the study area The mountain and sub-mountain climate is characterized by large contrasts within the local climate, which is rather cold, with a considerable amount of rainfall. Floods, occasionally disastrous, occur twice a year (spring and summer). The rivers are fed by rain, ground water and snow (Kondracki, 2000). The long-term annual average total precipitation is 846.63 mm (2011–2014). Total precipitation in the individual years in the study was 786.8 mm (2011), 813.8 mm (2012), 837.8 mm (2013), and 948.1 mm (2014). In the annual cycle, the highest total precipitation is typical of the summer and the lowest in the winter. The highest monthly total precipitation in the test period was 117
Catena 164 (2018) 116–124
W. Halecki et al.
was adapted to Polish conditions. With regard to the climate in the Polish Carpathians, where the winters are usually frosty and snowy, for the first month after last frost and the increase of temperature above zero degrees, R was calculated as a sum of rainfall erosivity plus the resulting surface runoff plus erosivity of melt runoff. For the months with temperatures > 0 °C, the melt runoff erosivity was omitted from the calculations. The R-factor is the product of kinetic energy of rainfall event (E) and its maximum 30-min intensity (I30) (Brown and Foster, 1987, Panagos et al., 2015a):
251.5 mm in July 2011 and the lowest 0.0 mm and 3.5 mm in October 2011 and January 2014, respectively. The highest daily precipitation was 100.7 mm on 15 May 2014. In the years 2011–2014, mean annual air temperature was 6.3 °C, and the hottest year, with mean annual temperature of 7.4 °C, was 2014. The lowest temperature of −25.8 °C was recorded on 3 February 2012 and the highest, 33.0 °C, on 8 July 2013. The snow starts to melt at the beginning of the March. The growing season starts around 10 April. The winter begins around 30 November. The mean number of days with light frost in the catchment area is between 120 and 130. Mean number of days with precipitation is 190, of which 70–80 are snowy days. Soils in the catchment area are diverse, depending on location on the slope. The soil cover in the Mątny stream catchment is dominated by loamy soils, including sandy clay loam, loam, silt loam, clay loam and sandy loam. Pedological conditions were identified by analysis of a 1:25,000 agricultural soil map and qualified in the respective groups according to USDA standards (Soil Survey Staff, 1975).
R=
V = e (LU ∙ Fcover )
Scale 1:5000, resolution 5m Scale 1:25,000
Digital map of soil
Digital map of land use
Meteorological data
– Institute of Meteorology and Water Management (2011–2014)
(5)
where: V — the vegetation retention (dimensionless), with V = 1 for bare land and V > 1 for managed land, LU — an empirical land-use parameter ranging from 1 to 10, and Fcover — a vegetation layer normalized within 0 to 1 range (in accordance with the principles of soil and water conservation). The LU values were derived from the CORINE Land Cover 2012 database (CLC, 2012). There are 33 soil type classes according to CLC, 2012 in Poland; of these, five classes are encountered in the Mątny stream catchment area. Agricultural lands include arable lands outside the reach of water supply systems, meadows and pastures as well as agricultural areas with a high share of natural vegetation. We compared deciduous forests, coniferous forests and mixed forests. Areas with scattered buildings have been excluded from these calculations; such settled area was not a very large factor in this site. The CLC classes were verified on the basis of on-site inspections on 20 July 2011, 25 July 2012, 15 June 2013 and 25 July 2014. The plant cover factor (Fcover) for grassland and forests was established to be 1 (Panagos et al. 2014a). Fcover for crops was established as a cultivated agri-environmental scheme in our study region (based on effective management, vegetation arrangement and individual features of plant growth). In addition, in the case of cereals (spring wheat and spring oats), the stubble field was left unplowed for the winter after the harvest. As regards arable land, potato harvesting was followed by the essential agrotechnical treatments (plowing, harrowing) and white mustard (Sinapis alba) was sown as a cover crop. Thus, the value of Fcover for arable land ranges from 0 to 1, depending on the plant development phase. The values of the plant vegetation cover are shown in Table 2.
Table 1 Categories, sources and description of G2 model for the study area.
Geodesic and Cartographic Documentation Center Institute of Soil Science and Plant Cultivation State Research Institute – Landsat – Orthophotomap – Copernicus Land (CORINE Land Cover 2012 PL) – Site control
(4)
2.3.3. Vegetation retention (V) The vegetation retention factor (V) is the degree to which the vegetation cover and management are expected to protect soil from erosion. In the G2 model, V was calculated based on Panagos et al. (2012, 2014a):
2.3.2. Rainfall erosivity (R) The rainfall-runoff erosivity factor for a specific time period – R –
DEM
−1
where: rain10 is precipitation of ≥10 mm; and day10 — number of days with precipitation above 10 mm.
where: E — the predicted soil loss from an area during a specific time period (Mg ha−1), R — the rainfall-runoff erosivity factor (MJ mm ha−1 h−1 (y/12)−1) for a specific time period which quantifies the impact of raindrop and runoff energy, V — the vegetation retention factor (dimensionless), S — the soil erodibility factor (Mg ha h MJ−1 ha−1 mm−1), T — the topography factor (dimensionless), and I — slope-intercept factor (dimensionless). Calculations for the model algorithm were carried out by means of analyses using ArcGIS 10.3.1 software. The spatial data were based on the Polish State Surveying Coordinate System (PUGW) 1992. Thematic layers and climatic data were developed using the sources specified in Table 1. The DEM basic thematic layer was developed on the basis of point coordinates (X, Y, Z) in a regular network with 1 m resolution, interpolated on the basis of a point cloud using LIDAR aviation laser scanning. DEM with a raster resolution of 5 × 5 m was used (obtained from the Geodesic and Cartographic Documentation Center in Poland).
Data description
(3)
l=1 k=1
EI30 = 7.05∙rain10 − 88.92∙day10
(2)
Source
lm
where: R — the rainfall-runoff erosivity factor (MJ mm ha h (y/ 12)−1), n — number of months covered by the data records, lj — number of erosive events of a given months m, and EI30 — the rainfall erosivity index of a single event k (MJ mm ha−1 h−1). Using this software module, these parameters were calculated for Obidowa station in Poland, for which the precipitation time series of 4 years (2011–2014) took into account precipitation of ≥10 mm (Loureiro and Coutinho, 2001; Suif et al., 2016). The event erosivity EI30 was defined as:
2.3.1. Model description The G2 model formula, expressed as a logical product, combines all the main factors, including natural and anthropogenic, that determine the nature and magnitude of soil erosion. Erosion intensity was assessed from the equation (Panagos et al., 2014a, 2014b):
Data type
n
∑ ∑ (EI30 )k −1
2.3. Model and data
R T E = ⎛ ⎞ ∙S∙ ⎛ ⎞ ⎝V ⎠ ⎝I⎠
1 n
– Multispectral satellite image – Scale 1:10,000 – Resolution 20 m – 20 July 2011 – 25 July 2012 – 15 June 2013 – 25 July 2014 – Weather station in Obidowa
2.3.4. Soil erodibility (S) Soil erodibility, S, was found from the formula proposed by Wischmeier and Smith (1978), modified by Renard et al. (1997) and 118
Catena 164 (2018) 116–124
W. Halecki et al.
image (which is 255 for a 8-bit system). Resolution in a model of the I-factor was between 1.0 and 1.58, with a negligible contribution of the parameter I > 1.42. The average Ifactor for the Mątny catchment was 1.14.
Table 2 The values of the plant vegetation cover. Land use
Spring oat/ common wheat Potatoes
Arable lands
Grasslands Forests
CORINE class
LU
Months
Parameter Fcover (−)
Parameter V (−)
211
6.50
211
6.50
IV V VI VII–III X–V VI VII VIII–IX I–XII I–XII
0.00 0.30 0.60 1.00 0.00 0.10 0.40 0.70 1.00 1.00
1.00 7.02 49.40 665.14 1.00 1,92 13.46 94.63 2980.96 8103.08
231 311/312/313
8.00 9.00
2.3.7. Relative risk index The contribution of the respective months to the annual value of water erosion was assessed from the relative risk index, as presented by Panagos et al. (2014a, 2014b) in the form:
STERIn =
1.14
10−4 ·(12–a) + 3.25·(b–2) + 2.5·(c–3) ⎤ ∙0.1317 ⎥ 100 ⎦ −1
−1
(6) −1
where: S — the soil erodibility factor (Mg ha h MJ ha mm ), M — the texture defined as percentage silt + fine sand fraction content multiplied by (100-clay fraction), a — the organic matter content (%); the a parameter is calculated as: a = 1.72 · orgC, where orgC is the organic carbon content in the soil layer (%), b — the soil texture class (very fine granular: 1, fine granular: 2, medium or coarse granular: 3 and very coarse granular: 4), and c — the profile permeability class (rapid: 1, moderate to rapid: 2, moderate: 3, slow to moderate: 4, slow: 5 and very slow: 6). In soils with a total content of silt and fine sand (grain size 0.002–0.2 mm) above 70%, the value of the soil erodibility factor S was found by means of a nomograph, developed by Wischmeier and Smith (1978). The characteristics required in the model were developed on the basis of soil samples collected from 43 sites in the catchment area. The values of the S factor for the various points ranged from 0.018 to 0.076 Mg ha h MJ−1 ha−1 mm−1. Lower values of the S factor were determined for sandy clay loam (0.018–0.034 Mg ha h MJ−1 ha−1 mm−1) and clay loam (0.022 Mg ha h MJ−1 ha−1 mm−1). For loam, the values were 0.036–0.068 Mg ha h MJ−1 ha−1 mm−1. Values exceeding 0.068 Mg ha h MJ−1 ha−1 mm−1 were recorded for silt soils.
2.3.8. Data processing Redundancy Analysis (RDA) was applied to show a common correlation pattern (Lepš and Šmilauer 2003) for analyzed features from different time periods (Fig. 5). This method might be seen as a promising contribution to validating the use of the main influential factors on water erosion and to document short-term soil stoniness within the model. RDA was used to quantify the importance of factors affecting soil erosion by water. Correlation matrices were selected for land use. The statistical procedure was computed by the program Canoco for Windows ver. 4.51. 3. Results 3.1. Investigation of short-term soil erosion The spatial distribution of yearly erosion risk classes in the investigated area using the G2 model in the respective years from 2011 to 2014 period is shown in a set of maps (Fig. 3). The eroded soil mass (Mg y−1) in the Mątny stream catchment area, as calculated using the G2 model for the respective years in the period 2011–2014 was as follows: 967.92 (2011), 494.87 (2012), 1418.32 (2013), and 4563.87 (2014). Total annual soil loss and water erosion for the years 2011–2014 was 12.7 Mg ha−1. The maximum eroded soil mass was 31.05 Mg per ha of area. The lowest total annual soil loss was 3.37 Mg ha−1, recorded in 2012. The calculated standard deviation for the multi-year period was 12.52 Mg ha−1, and the variability coefficient was 99%. In the year 2011, the lowest value for total soil eroded mass was obtained in October (16.94 Mg). In 2012 the highest value was 238.90 Mg in June. Total soil eroded mass in the other months was similar (March: 20.50 Mg, August: 18.87 Mg, September: 16.26 Mg, October and November: 10.99 Mg). Total soil eroded mass due to water erosion in the summer months (June–August) was 88.13% of the total annual soil loss. In the autumn and winter periods, it accounted for 7.73 and 4.15% of the total annual soil loss, respectively. In the year 2013, the total soil eroded mass share of the months May–December in the total annual soil loss ranged from 0.46% (December) to 57.08% (June). For the year 2014, total soil eroded mass in the spring months (March–May) was 81.62% of the total annual soil loss. In the summer (June–August) and autumn (September–November) the total soil eroded mass was 16.74% and 1.64%, respectively, of total annual soil loss. Data of the relative risk index (STERIn) is presented in Table 3. The relative risk index values for June, July and August were 27.89, 15.09 and 3.48%, respectively. RDA detected that vegetation retention parameter was highly
2.3.5. Topographic influence (T) To estimate the effects of topography on the erosion risk, an equation developed by Moore and Burch (1986) and proposed for use in erosion modeling:
sin β ⎞1.3 A 0.4 LS = T = ⎛ s ⎞ ∙ ⎛ ⎝ 22.13 ⎠ ⎝ 0.0896 ⎠
(7)
where: T — the topographic influence (≥0, dimensionless), As — the flow accumulation (m), and β — the slope (rad). The respective equation parameters were found using the ArcMap 10.3.1 software. The values of the T parameter ranged from 0.00 to 38.79 (the mean value was 4.19). 2.3.6. Slope intercept (I) The slope intercept (I) was estimated by applying a Sobel filter on the high resolution satellite image mosaic. The equation for the computation of the I factor was compared with that proposed by Panagos et al. (2012) in order to include the I-factor in the denominator in the previous equation (Karydas et al., 2015):
I=1+
Sf DNmax
(9)
where: STERIn — relative risk index [–], En — water erosion for the respective months, and E — annual water erosion. The extent of soil erosion was determined as per an erosion risk scale proposed by Marks et al. (1989), which divides erosion risk into six classes: no risk: < 1, very low risk: 1–5; low risk: 5–10, medium risk: 10–15, high risk: 15–30, and very high risk: > 30, expressed in Mg ha−1.
presented by Panagos et al. (2012):
2.1·M S=⎡ ⎢ ⎣
En E
(8)
where: I — the slope intercept [–], Sf — value of digital image mosaic as generated using the Sobel filter with a 3 × 3 mask, adopting values from 0 to DNmax, and DNmax — maximum value of the digital Sobel 119
Catena 164 (2018) 116–124
W. Halecki et al.
Table 3 Total annual soil loss and STERIn in 2011–2014 period. Month/year
January February March April May June July August September October November December
STERIn (%)
Total annual soil loss (Mg) 2011
2012
2013
2014
2011
2012
2013
2014
– – – – 501.25 269.93 145.10 33.71 – 16.94 – –
0.02 20.50 – – – 238.9 178.33 18.87 16.26 10.99 10.99 –
– – – – 477.22 809.57 29.83 14.70 17.20 34.79 28.50 6.52
– – – 56.13 3669.19 606.84 137.06 19.89 41.86 41.86 – –
– – – – 51.79 27.89 15.09 3.48 – 1.75 – –
0.00 4.14 – – – 48.28 36.04 3.81 3.29 2.22 2.22 –
– – – – 33.65 57.08 2.10 1.04 1.21 2.45 2.01 0.46
– – – 1.23 80.40 13.30 3.00 0.43 0.92 0.72 – –
Fig. 2. Distribution of the rainfall erosivity (R factor) for the years 2011–2014.
4. Discussion
correlated with forest, while slope intercept and event erosivity parameters were negatively related with grassland as well as arable area (Fig. 5). The estimated R-factor values (Fig. 2) were characterized by strong seasonality, with the highest regime coefficients occurring from May to July (e.g. May 2013: 110.41 MJ mm ha−1 h−1; May 2014: 838.90 MJ mm ha−1 h−1) and the lowest (almost zero) occurring in March. The mean estimated R-factor for the Mątny catchment was 92.5 MJ mm ha−1 h−1. The mean annual value of the parameter R (MJ mm ha−1 h−1 y−1) for the catchment was 94.68 (2011), 86.34 (2012), 106.02 (2013), and 234.75 (2014).
4.1. Modeling of water erosion at the catchment scale The melt runoff erosivity Rs was found in accordance with the USLE method (Wischmeier and Smith, 1978), assuming that its value for the first month in a year with temperatures > 0 °C (March) is 1/10 of the total precipitation layer, as expressed in millimeters of water column in the preceding period of time with temperatures < 0 °C. The calculated values of Rs were 5.6 mm (2011), 17.2 mm (2012), 16.2 mm (2013), and 4.2 mm (2014). In Poland, the method for the Rs factor assessment was also used e.g. by Banasik and Górski (1990), and Kruk et al. (2016). The monthly distribution of rainfall-runoff erosivity factor R for a specific time period is shown in Fig. 2. The value of the R factor seems closely correlated with total annual precipitation. The relationship is described by the power function, fitted with the correlation coefficient r = 0.76. The effect of kinetic energy of rainfall EI30 on the value of the R factor was characterized by the correlation coefficient r = 0.91. Similar seasonal patterns were reported by Świętochowicz (2013) for the Limanowa station in Beskid Wyspowy (amounting to 113.5 (MJ mm ha−1 h−1 y−1) and a model fit (correlation coefficient — R) equal to 0.75) where winter plowing was used. According to the ESDAC program (European Soil Data Centre), which is based on the RUSLE2015 program and the EIONET-SOIL project (Joint Research Centre, European Commission), the Mątny stream catchment area generates a total soil eroded mass of 5–10 Mg ha−1 y−1. Kowalczyk and Twardy (2012) indicated the role of the soil and water conservation and the effects of inappropriate management in the
3.2. Erosion intensity Erosion intensity distribution in the respective months was analyzed against the land-use (Fig. 4). The highest total annual soil losses were recorded in 2014. For spring oats, the losses in 2014 were 9.27 Mg ha−1, for potatoes 650.78 Mg ha−1, for grassland 3.24 Mg ha−1, and for forests 0.06 Mg ha−1. With reference to land cover, the highest soil losses were recorded in the spring and summer months on arable lands where root plants were cultivated. The values ranged from 68.31% (2012) to 90.12% (2014) of the total annual soil loss. The contribution of grassland to the annual water erosion ranged from 31.11% (2012) to 7.68% (2014) of the total annual soil loss. For the years in this study, the total soil eroded mass in forest covered areas was negligible, not higher than 0.1% of its total value.
120
Catena 164 (2018) 116–124
W. Halecki et al.
Fig. 3. Variability of soil loss (detachment soil participles through transport and surface-runoff) caused by water erosion in consecutive years.
catchment of the Bystrzyca Dusznicka (197.7 km2) in the south-west of Poland. In this study a mean total annual soil loss was 900.35 Mg km−2. Panagos et al. (2014a) studied soil erosion using the G2 method on Crete with precipitation data from 24 weather stations for the multiyear period 1969–1979. Water erosion assessment led to soil loss underestimation, especially for croplands and grasslands. It is, therefore, still of prime importance to improve the newly developed method. In our erosion model, we integrated all factors connected with total soil eroded mass. In further research, the G2 model can be used to predict soil erosion even under natural vegetation. Karydas et al. (2015) performed complex studies on actual soil erosion using the G2 model for various types of crops. Mean monthly values of potential erosion intensity are essential for appraisal of the potential risk of water erosion. This is one of the principal advantages of the G2 model. Our study showed that soil loss assessment is substantial to control erosion by water. We suggest that soil and water conservation is key in reducing soil erosion in arable areas. However, erosion
intensity can be increased in the crops such as potatoes, and nearby native evergreen forests in erosion-prone area. 4.2. Consequences of land-use changes for soil erosion by water It is crucial to design appropriate efficient tools for water erosion and land management (Bosco et al., 2015; Jha et al., 2015). Obviously, predicting runoff distribution has been presented in the literature where soil erosion intensively occurs and clearly points towards the importance of empirical and theoretical models (White et al., 2011; Tilahun et al., 2014; Rathjens et al., 2015; Hoang et al., 2017). Notoriously, in the European Union countries, mean values were very close to the 2.46 t ha1 yr1 analytically estimated in erosion-prone lands. In our study, mean annual soil erosion was between 3.37 and 31.04 t ha1 yr1. To mitigate land degradation processes, there is a need to provide the reduction of soil erosion by water based on support practices — P-factor (Panagos et al., 2015a,b). Moreover, spatial 121
Catena 164 (2018) 116–124
W. Halecki et al.
Fig. 4. Temporal trends (monthly mean) of R (rainfall erosivity), V (vegetation retention) and E (erosion intensity) parameters for the years 2011–2014.
Table 4 R-value according to our method (minimum, maximum and mean in the period 2011–2014) and Ballabio et al. (2017). Month
Minimum R
Maximum R
Mean R 2011–2014
Ballabio et al. (2017)
January February March April May June July August September October November December
0 0 0 0 110.41 103.89 61.16 81.6 0 52.08 0 0
0.1 97.2 0 56.92 848.97 352.05 365.63 145.76 181.02 164.88 135.09 30.93
0 24 (+3↑) 0 14 269 (+168↑) 209 252 94 (−62↓) 81 112 47 (+14↑) 8
– < 21 < 21 < 21 66–99 > 156 > 157 > 156 66–99 99–158 21–33 < 21
↑ means overestimated value; ↓ illustrate underestimated value.
analysis of soil features is helpful in determination of factors affecting the saturation-excess runoff type in mountain areas, particularly for sustainable agricultural practices and land use (Halecki et al., 2016; Halecki et al., 2017). In Polish conditions water erosion is highly dependent on recent vegetative phase. In a case of oat and wheat the value of V parameter attains the minimum on April, and the maximum occurs between July and March. This tendency is the same for potatoes, however, in another periods of total soil eroded mass. Monthly erosive periods are not influential on grasslands and forests (Table 4). Our study revealed that land-cover corresponded with soil erodibility, rainfall erosivity and vegetation retention in a profound way (Fig. 5). This is highly probable because, nowadays, not all farmers tend to implement erosion-limiting treatments, such as leaving unplowed stubble field over the winter, or the use of cover crops.
Fig. 5. RDA plot. Solid squares represent arable area, rectangles indicate grassland and diamonds designate — forest zone. Eigenvectors (arrows) stand for factors affecting soil erosion by water in different land use. The two components of the RDA analysis explained 72.8% of a total variability. Explained variance for the first axis was 40.8% and for the second one was 32.0%.
for the first time under conditions prevalent in Poland. Undoubtedly the advantage of the model is the possibility of assessing total soil eroded mass over monthly intervals. It enables the determination of the influence of plant growing stages on the V parameter shaping. In the work, details from the CLC2012 database, based on satellite orthophotomaps on field observation, was demonstrated. A limitation on the model is the resolution of maps from CORINE Land Cover database. The G2 model introduces a new quality in the assessment of soil erosion intensity. The monthly approach of the model using various plant stages was important for soil and water conservation. The study verified the stochastic method of estimating soil erosion under different regimes of sediment transport and land use. The presented results indicated the relative consequence of land use. The G2 model, as we conjecture from our preliminary investigations, seems to be dominated by the creation of short-term erosive events and the detachment of soil material. Our calculations have also showed the important role of estimating conservation measures against water erosion, and focused on soil erodibility. We suggest that the G2 model is a useful tool for surface water erosion appraisal such as rain drop and rill erosion on small-scale in short time intervals. Long-term data series should be implemented in
5. Conclusion The monthly approach to the G2 model enabled an identification of the months that generated the highest total soil eroded mass in Polish mountainous weather conditions. Redundancy analysis based on landuse types displayed factors affecting soil erosion by water. The highest total soil eroded mass was recorded in the spring and summer months on arable land planted with potatoes. The annual value of water erosion in the catchment was also found to be negatively affected by the presence of grassland, in addition to cultivated plants. This is associated with sheep- and cattle grazing in the area. In contrast, total soil eroded mass was negligible in the forests. Here, the G2 model has been verified 122
Catena 164 (2018) 116–124
W. Halecki et al.
further investigations to improve modeling and calibration schemes by simulating high rates of soil erosion (rill erosion) on a larger scale.
GIS in a Shivalik subwatershed. J. Earth Syst. Sci. 122 (2), 389–398. Laflen, J.M., Moldenhauer, W.C., 2003. Pioneering soil erosion prediction: the USLE story. In: Word Association of Soil and Water Conservation Special Publication. 1. pp. 54. Lee, G.S., Choi, I.H., 2010. Scaling effect for the quantification of soil loss using GIS spatial analysis. KSCE J. Civ. Eng. 14 (6), 897–904. http://dx.doi.org/10.1007/ s12205-010-0878-x. Lepš, J., Šmilauer, P., 2003. Multivariate Analysis of Ecological Data Using CANOCO. Cambridge University Press, New York. Loureiro, N.S., Coutinho, M.A., 2001. A new procedure to estimate the RUSLE EI30 index, based on monthly rainfall data and applied to the Algarve Region Portugal. J. Hydrol. 250, 12–18. Mao, D., Cherkauer, K.A., 2009. Impact of land-use change on hydrologic responses in the Great Lakes region. J. Hydrol. 374, 71–82. http://dx.doi.org/10.1016/j.jhydrol. 2009.06.016. Marks, R., Müller, M.J., Leser, H., Klink, H.J., 1989. Anleitung zur Bewertung des Leistungsvermögens des Landschaftshaushaltes (BA LVL). In: Zentralausschuß für deutsche Landeskunde. Forschungen zur Deutschen Landeskunde. 229. Mhangara, P., Kakembo, V., Lim, K., 2012. Soil erosion risk assessment of the Keiskamma catchment, South Africa using GIS and remote sensing. Environ. Earth Sci. 65 (7), 2087–2102. Moore, I.D., Burch, G.J., 1986. Sediment transport catchment of sheet and rill flow: application of unit stream power theory. Water Resour. Res. 27 (8), 1350–1360. Nadal-Romero, E., Regüés, D., Latron, J., 2008. Relationships among rainfall, runoff, and suspended sediment in a small catchment. Catena 74 (2), 127–136. http://dx.doi. org/10.1016/j.catena.2008.03.014. Panagos, P., Karydas, C.G., Gitas, I.Z., Montanarella, L., 2012. Monthly soil erosion monitoring based on remotely sensed biophysical parameters: a case study in Strymonas river basin towards a functional pan-European service. Int. J. Digital Earth 5 (6), 461–487. https://doi.org/10.1080/17538947.2011.587897. Panagos, P., Karydas, C.G., Ballabio, C., Gitas, I.Z., 2014a. Seasonal monitoring of soil erosion at regional scale: an application of the G2 model in Crete focusing on agricultural land uses. Int. J. Appl. Earth Obs. Geoinf. 27PB, 147–155. http://dx.doi.org/ 10.1016/j.jag.2013.09.012. Panagos, P., Meusburger, K., Ballabio, C., Borrelli, P., Alewell, C., 2014b. Soil erodibility in Europe: a high-resolution dataset based on LUCAS. Sci. Total Environ. 479 –480, 189–200. http://dx.doi.org/10.1016/j.scitotenv.2014.02.010. Panagos, P., Ballabio, C., Borrelli, P., Meusburger, K., Klik, A., Rousseva, S., Tadić, M.P., Michhaelides, S., Hrabalíková, M., Olsen, P., Aalto, J., Lakatos, M., Rymszewicz, A., Dumitrescu, A., Beguería, S., Alewell, C., 2015a. Rainfall erosivity in Europe. Sci. Total Environ. 511, 801–814. http://dx.doi.org/10.1016/j.scitotenv.2015.01.008. Panagos, P., Borrelli, P., Poesen, J., Ballabio, C., Lugato, E., Meusburger, K., Montanarella, L., Alewell, C., 2015b. The new assessment of soil loss by water erosion in Europe. Environ. Sci. Pol. 54, 438–444. Panagos, P., Borrelli, P., Meusburger, K., van der Zanden, E.H., Poesen, J., Alewell, C., 2015c. Modelling the effect of support practices (P-factor) on the reduction of soil erosion by water at European Scale. Environ. Sci. Pol. 51, 23–34. Park, S., Oh, C., Jeon, S., Jung, H., Choi, C., 2011. Soil erosion risk in Korean watersheds, assessed using the revised universal soil erosion equation. J. Hydrol. 399, 263–273. http://dx.doi.org/10.1016/j.hydrol.2011.01.004. Perović, V., Životić, L., Kadović, R., Đorđević, A., Jaramaz, D., Mrvić, V., Todorovic, M., 2013. Spatial modelling of soil erosion potential in a mountainous watershed of South-eastern Serbia. Environ. Earth Sci. 68, 115–128. Podwojewski, P., Orange, D., Jouqert, P., Valentin, C., Nguyen, C.T., 2008. Land-use impacts on surface runoff and soil detachment within agricultural sloping lands in Northern Vietnam. Catena 74, 109–118. http://dx.doi.org/10.1016/j.catena.2008. 03.013. Rathjens, H., Oppelt, N., Bosch, D.D., Arnold, J.G., Volk, M., 2015. Development of a gridbased version of the SWAT landscape model. Hydrol. Process. 29, 900–914. Renard, K.G., Foster, G.R., Weesies, G.A., McCool, D.K., Yoder, D.C., 1997. Predicting soil erosion by water: a guide to conservation planning with the Revised Universal Soil Loss Equation (RUSLE). In: Agriculture Handbook. 73. US Department of Agriculture, Washington, DC, pp. 1–251. Smith, S.J., Wiliams, J.R., Menzel, R.G., Coleman, G.A., 1984. Prediction of sediment yield from southern plains grasslands with the Modified Universal Soil Loss Equation. J. Range Manag. 37, 295–297. Soil Survey Staff, 1975. Soil Taxonomy. USDA, Washington, pp. 17–754. Suif, Z., Fleifle, A., Yishimura, C., Saavedra, O., 2016. Spatio-temporal patterns of soil erosion and suspended sediment dynamics in the Mekong River Basin. Sci. Total Environ. 568, 933–945. Świętochowicz, J., 2013. Mean annual R factor for Łazy near Bochnia meteorological station (Carpathian Foothills climatic region). Landf. Anal. 24, 85–95. Tilahun, S.A., Guzman, C.D., Zegeye, A.D., Ayana, E.K., Collick, A.S., Yitaferu, B., Steenhuis, T.S., 2014. Spatial and Temporal Patterns of Soil Erosion in the Semihumid Ethiopian Highlands: A Case Study of Debre Mawi Watershed. Springer, Nile River Basin, pp. 149–163. Ustun, B., 2008. Soil erosion modelling by using GIS and remote sensing: a case study Ganos Mountain. Int. Arch. Photogramm. Remote. Sens. Spat. Inf. Sci. 37, 681–1684. Verstraete, G., Poesen, J., 2001. Factors controlling sediment yield from small intensively cultivated catchments in a temperate humid climate. Geomorphology 40 (1), 123–144. http://dx.doi.org/10.1016/S0169-555X(01)00040-X. White, E.D., Easton, Z.M., Fuka, D.R., Collick, A.S., Adgo, E., McCartney, M., Steenhuis, T.S., 2011. Development and application of a physically based landscape water balance in the SWAT model. Hydrol. Process. 25, 915–925. http://dx.doi.org/10. 1002/hyp.7876. Wischmeier, W.H., Smith, D.D., 1978. Predicting rainfall erosion losses — a guide to
Acknowledgements This research was financed by the Ministry of Science and Higher Education of the Republic of Poland. The study was also supported by the Projects No. DS-3331/KMiKS/2011-2017. References Bagarello, V., Ferro, V., Pampalone, V., 2015. A new version of the USLE-MM for predicting bare plot soil loss at the Sparacia (South Italy) experimental site. Hydrol. Process. 29, 420–4219. http://dx.doi.org/10.1002/hyp.10486. Bahadur, K.C.K., 2009. Mapping soil erosion susceptibility using remote sensing and GIS: a case of the Upper Nam Wa Watershed, Nan Province, Thailand. Environ. Geol. 57, 695–705. http://dx.doi.org/10.1007/s00254-008-1348-3. Ballabio, C., Borrelli, P., Spinoni, J., Meusburger, K., Michaelides, S., Beguería, S., Klik, A., Petan, S., Janeček, M., Olsen, P., Aalto, J., Lakatos, M., Rymszewicz, A., Dumitrescu, A., Tadic, M.P., Diodato, N., Kostalova, J., Rousseva, S., Banasik, K., Alewell, C., Panagos, P., 2017. Mapping monthly rainfall erosivity in Europe. Sci. Total Environ. 579, s.1298–1315. Banasik, K., Górski, D., 1990. Wyznaczanie erozyjności deszczy do uniwersalnego równania strat glebowych. In: Zeszyty Naukowe AR Wrocław, Melioracje. 34. pp. 103–109. Bosco, C., de Rigo, D., Dewitte, O., Poesen, J., Panagos, P., 2015. Modelling soil erosion at European scale: towards harmonization and reproducibility. Nat. Hazards Earth Syst. Sci. 15, 225–245. Brown, L.C., Foster, G.R., 1987. Storm erosivity using idealized intensity distributions. Trans. ASAE 30, 379–386. Burrough, P.A., 2001. GIS and geostatistics: essential partners for spatial analysis. Environ. Ecol. Stat. 8, 361–377. http://dx.doi.org/10.1023/A:1012734519752. Cârdei, P., 2010. The dimensional analysis of the USLE — MUSLE soil erosion model. Proc. Rom. Acad., Ser. B 3, 249–253. CLC, 2012. CORINE Land Cover. http://inspire.gios.gov.pl/portal/index.php?profile= 29542&projection=EPSG%3A2180. Dabral, P.P., Baithuri, A., Pandey, A., 2008. Soil erosion assessment in a hilly catchment of North Eastern India using USLE, GIS and remote sensing. Water Resour. Manag. 22, 1783–1798. http://dx.doi.org/10.1007/s11269-008-9253-9. Flanagan, D.C., Lafen, J.M., Meyer, L.D., 2003. American Society of Agricultural Engineers Historic Landmark Dedication; Honoring the Universal Soil Loss Equation. Agric. & Biol. Eng., Purdue University. Halecki, W., Młyński, D., Ryczek, M., Kruk, E., Lis, J., 2016. Zmienność przestrzenna wilgotności i gęstości objętościowej gleby w zlewni potoku górskiego Mątny w Gorcach. In: EPISTEME. 30. pp. 347–358 ISSN 1895-4421. Halecki, W., Młyński, D., Ryczek, M., Kruk, E., Radecki-Pawlik, A., 2017. The application of Artificial Neural Network (ANN) to assessment of soil salinity and temperature variability in agricultural areas of a mountain catchment. Pol. J. Environ. Stud. 6, 2545–2554. Hoang, L., Schneiderman, E.M., Moore, K.E.B., Mukundan, R., Owens, E.M., Steenhuis, T.S., 2017. Predicting saturation-excess runoff distribution with a lumped hillslope model: SWAT-HS. Hydrol. Process. 31, 2226–2243. http://dx.doi.org/10.1002/hyp. 11179. Jha, A., Schkade, U., Kirchner, G., 2015. Estimating short term soil erosion rates after single and multiple rainfall events by modelling the vertical distribution of cosmogenic 7Be in soils. Geoderma 243–244, 149–156. Karydas, C.G., Panagos, P., 2016. Modelling monthly soil losses and sediment yields in Cyprus. Int. J. Digital Earth 9 (8), 766–778. Karydas, C.G., Panagos, P., 2018. The G2 erosion model: an algorithm for month-time step assessments. Environ. Res. 161, 256–267. http://dx.doi.org/10.1016/j.envres. 2017.11.010. Karydas, C.G., Panagos, P., Gitas, I.Z., 2014. A classification of water erosion models according to their geospatial characteristics. Int. J. Digital Earth 7 (3), 229–250. http://dx.doi.org/10.1080/17538947.2012.671380. Karydas, C.G., Zdruli, P., Koci, S., Sallaku, F., 2015. Monthly time-step erosion risk monitoring of Ishmi-Erzeni watershed, Albania, using the G2 model. Environ. Model. Assess. 20 (6), 657–672. http://dx.doi.org/10.1007/s10666-015-9455-5. Kinnell, P.I.A., 2016. Comparison between the USLE, the USLE-M and replicate plots to model rainfall erosion on bare fallow area. Catena 145, 39–464b. http://dx.doi.org/ 10.1016/j.catena.2016.05.017. Kinnell, P.I.A., Risse, L.M., 1998. USLE-M: empirical modeling rainfall erosion through runoff and sediment concentration. Soil Sci. Soc. Am. J. 62, 1667–1672. Kondracki, J., 2000. Geografia regionalna Polski. PWN, Warszawa. Kowalczyk, A., Twardy, S., 2012. Wielkość erozji wodnej obliczona metodą USLE. WodaŚrodowisko-Obszary Wiejskie 12 (37), 83–92. Kruk, E., 2017. Influence of daily precipitation on yield of eroded soil in mountain basin using the Musle model. Acta. Sci. Pol. Form. Circum. 16 (2), 147–158. http://dx.doi. org/10.15576/ASP.FC/2017.16.2.147. Kruk, E., Ryczek, M., Malec, M., 2016. Porównanie wielkości strat glebowych wyznaczonych modelami USLE i RUSLE w zlewni górskiej. Wybrane aspekty badawcze. In: Aspekty teoretyczno-praktyczne badań naukowych. 3. pp. 55–65. Kumar, S., Kushwaha, S.P.S., 2013. Modeling soil erosion risk based on RUSLE-3D using
123
Catena 164 (2018) 116–124
W. Halecki et al.
Korçe region, Albania. Earth Sci. Inf. 9 (4), 481–495. Zhang, Y., Degroote, J., Wolter, C., Sugumaran, R., 2009. Integration of modified universal soil loss equation (MUSLE) into a gis framework to asses soil erosion risk. Land Degrad. Dev. 20, 84–91. http://dx.doi.org/10.1002/ldr.893. Żmuda, R., Sasik, J., Szewrański, S., 2005. Analiza potrzeb zmian zagospodarowania przestrzennego wzgórz trzebnickich w aspekcie ochrony przed erozją wodną gleb. Acta Agrophysica 5 (1), 229–237.
conservation planning. In: Supersedes Agriculture Handbook No. 282, pp. 4–11 Washington. Zabaleda, A., Martinez, M., Mriarte, J.A., Antiguedad, I., 2007. Factors controlling suspended sediment yield during runoff events in small headwater catchments of the Basque country. Catena 71 (1), 179–190. http://dx.doi.org/10.1016/j.catena.2006. 06.007. Zdruli, P., Karydas, C.G., Dedaj, K., Salillari, I., Cela, F., Lushaj, S., Panagos, P., 2016. High resolution spatiotemporal analysis of erosion risk per land cover category in
124
Contents lists available at ScienceDirect
Catena journal homepage: www.elsevier.com/locate/catena
Evaluation of water erosion at a mountain catchment in Poland using the G2 model ⁎
⁎
Wiktor Haleckia, , Edyta Krukb, , Marek Ryczekb,
T
⁎
a
Department of Biometry and Forest Productivity, Faculty of Forestry, University of Agriculture in Krakow, Al. 29 Listopada 46, 31-425 Kraków, Poland Department of Land Reclamation and Environmental Development, Faculty of Environmental Engineering and Land Surveying, University of Agriculture in Krakow, 30059 Kraków, Al. Mickiewicza 24-28, Poland b
A R T I C L E I N F O
A B S T R A C T
Keywords: Erosion risk assessment G2 model Land use Poland Soil loss Surface-runoff
The Western Carpathians region in southern Poland is characterized by high erosion risk due to steep slopes, flysch formation and intense precipitation with frequent storm events. The G2 model, based on the principles of the Universal Soil Loss Equation (USLE), was used to investigate soil erosion assessment, except that the yearly rainfall erosivity factor was substituted for by the monthly one. The plant cover factor was determined based on the CORINE land cover 2012 database and field observations of vegetation stages. Slope intercept was estimated by applying a Sobel filter. Terrain properties were calculated from a 5 m DEM of the area. Modeling investigations were carried out in the agricultural basin of the mountain catchment (1.47 km2) in the years 2011–2014. Rainfall data were collected from weather stations, and soil properties were measured in 43 locations. The G2 model estimated total annual soil loss as between 3.37 Mg ha−1 (2012) and 31.05 Mg ha−1 (2014). The erosive events that contributed most to yearly erosion occurred in May (2014: 80.40% of yearly total) and June (2013: 57.08%). Redundancy analysis based on land-use types provided factors affecting soil erosion by water. In conclusion, the G2 model was useful in erosion estimation in a steep-sloped agricultural basin with a variable hydrological regime.
1. Introduction Water erosion of soil is a complex process during which fertile topsoil is detached, carried away and deposited in another location (Ballabio et al., 2017). It causes leaching of nutrients, terrain surface deformations and deterioration of water quality in catchments as well as silting of water structures and water supply and drainage systems (Żmuda et al., 2005). Even slight erosion negatively affects farming conditions and yield, impedes agrotechnical treatments, and may exclude the entire affected area from agriculture (Halecki et al., 2016). The intensity of erosion events is shaped by physiographic and hydrological factors prevailing in a specific catchment area (Verstraete and Poesen, 2001; Zabaleda et al., 2007), and by land use and plant cover (Podwojewski et al., 2008; Mao and Cherkauer, 2009). It also highly depends on geological factors, soil type, and climate (Nadal-Romero et al., 2008). The accuracy of soil loss evaluation with models depends largely on how the model parameters describe important characteristics of the catchment. Many researchers focus on the assessment of the parameters provided by the Geographic Information System (GIS) and remote
⁎
sensing data. These techniques are useful for quick evaluation of the spatial distribution of erosion within large areas, and can cover remote areas where no measurements are actually conducted (Dabral et al., 2008; Bahadur, 2009). The parameters estimated with these techniques do not differ much from those obtained from field measurements (Lee and Choi, 2010). Additionally, new geostatistical methods allow for development of maps with predicted data values that previously have had to be developed from highly time-consuming and expensive point observations (Burrough, 2001). Telemetry combined with GIS techniques and geostatistical methods enable the implementation of holistic and integrated visions of sustainable development. The most important climatic factor is precipitation, and its frequency, duration, and intensity are critical for erosion in a catchment. Water erosion processes play an important role in terrain shaping causing transformations that under natural conditions result in formation of new forms of relief. In Poland, these processes are particularly visible in mountain and foothill areas. The literature provides many generations of erosion models describing water runoff and erosion events that differ in their limiting factor(s). The most popular and widely used is the Universal Soil Loss
Corresponding authors. E-mail addresses: [email protected] (W. Halecki), [email protected] (E. Kruk), [email protected] (M. Ryczek).
https://doi.org/10.1016/j.catena.2018.01.014 Received 21 August 2017; Received in revised form 9 January 2018; Accepted 15 January 2018 0341-8162/ © 2018 Elsevier B.V. All rights reserved.
Catena 164 (2018) 116–124
W. Halecki et al.
Fig. 1. Location of the investigated catchment of the Mątny stream.
generation among all the available water erosion models of the USLE family. There has been a lack of investigation on its use in Polish conditions. Therefore, the objectives of this study were (1) to apply the G2 model to quantitative assessment of water erosion in an agricultural mountain catchment and (2) to evaluate the effect of soil and water conservation (SWC) on total soil eroded mass in a highly erosion-prone landscape.
Equation model (USLE) developed in a series of papers by Wischmeier and Smith (1978); Flanagan et al. (2003); Laflen and Moldenhauer (2003); Perović et al. (2013); and Kruk et al. (2016). The model formula, expressed as a logical product that combines all main natural and anthropogenic factors shaping the type and extent of soil erosion is as follows:
A = RKLSCP
(1)
2. Materials and methods
where: A — mean annual weight of eroded soil per unit area (Mg ha−1 y−1), R — mean annual erosivity of rainfall and runoff (MJ mm ha−1 h−1 y−1), K — soil erodibility (Mg ha h MJ−1 ha−1 mm−1), L — slope length coefficient (dimensionless), S — slope steepness coefficient (dimensionless), C — coefficient determining the type of crops and land use (dimensionless), and P — soil and water conservation (dimensionless). The USLE model is used to estimate total annual soil loss within a multi-year period. It has been constantly modified and amended as research progress is made and the availability and quality of software are improved. The following modifications are worth mentioning: MUSLE (Smith et al., 1984; Zhang et al., 2009; Cârdei, 2010, Kruk, 2017), RUSLE (Park et al., 2011; Mhangara et al., 2012; Kumar and Kushwaha, 2013), USLE-M (Kinnell and Risse, 1998; Kinnell, 2016), and USLE-MM (Bagarello et al., 2015). In 2012, G2, a new erosion model based on the USLE structure was developed at the Aristotle University of Thessaloniki and Joint Research Centre of the European Commission under the GEOLAND2 project. The model formula was elaborated by Panagos et al. (2012, 2014b, 2015a), Karydas et al. (2014, 2015), Karydas and Panagos (2016) and Zdruli et al. (2016). It is used to estimate the erosion extent on a monthly basis. Several published studies of G2 implementation, e.g. in the cross-borders basin of the Strymonas/Struma river (Greece and Bulgaria), in the basins of the Ishmi-Erzeni river and Korce region (Albania), in the Mediterranean islands of Crete (Greece) and Cyprus, have yielded realistic results (Panagos et al. 2014a; Panagos et al. 2015a,c, Karydas and Panagos, 2016, 2018, Zdruli et al. 2016). The choice of erosion modeling method usually depends on availability of proper data and the purpose of the simulation. An important advantage of the presented models is their simplicity and low requirements concerning the input data as compared with other erosion models such as SWAT — Soil and Water Assessment Tool (Ustun, 2008). The G2 model can assess soil loss in cropland and forest-dominated land uses. We hypothesized that systematically enhancing soil conservation practices might substantially reduce soil loss in consecutive years. We decided to use the G2 model because it was the newest
2.1. Study area description The study was conducted in the Outer Western Carpathians, in the southern region of Małopolska Province (Fig. 1). The terrain comprises low and medium height mountains with peak heights ranging from 617.6 m a.s.l. to 732.0 m a.s.l. The lowest point is situated 490.0 m a.s.l. The mean height of the catchment is 582.66 m a.s.l. The slope distribution is as follows: < 5%, 4.08%; 5–10%, 18.37%; 10–18%, 45.58%; 18–27%, 21.09%; and > 27%, 10.88%. The weighted mean slope for the entire catchment is 16.28%. Study area was characterized by a deep or relatively deep soil profile. The length of the main watercourse is 2.37 km and its average slope 5.7%. The river network density is 2.96 km km−2. The Mątny stream (1.47 km2) flows into the Mszanka river in the Skiby hamlet at E 20°9′2.35″, N 49°37′30.52″. The catchment land use structure is dominated by grassland (73.5%); an arable land cover of 14.3%, including spring oat (Avena sativa) — 7.3%, potatoes (Solanum tuberosum) — 4.3%, common wheat (Triticum aestivum) — 2.7%; forests account for 9.5%; and urban areas 2.7%. The catchment area is cut by a network of dirt roads. Most of them are deeply furrowed and tend to transform into water-carrying streams during and after rain events. 2.2. Climatic conditions and soil of the study area The mountain and sub-mountain climate is characterized by large contrasts within the local climate, which is rather cold, with a considerable amount of rainfall. Floods, occasionally disastrous, occur twice a year (spring and summer). The rivers are fed by rain, ground water and snow (Kondracki, 2000). The long-term annual average total precipitation is 846.63 mm (2011–2014). Total precipitation in the individual years in the study was 786.8 mm (2011), 813.8 mm (2012), 837.8 mm (2013), and 948.1 mm (2014). In the annual cycle, the highest total precipitation is typical of the summer and the lowest in the winter. The highest monthly total precipitation in the test period was 117
Catena 164 (2018) 116–124
W. Halecki et al.
was adapted to Polish conditions. With regard to the climate in the Polish Carpathians, where the winters are usually frosty and snowy, for the first month after last frost and the increase of temperature above zero degrees, R was calculated as a sum of rainfall erosivity plus the resulting surface runoff plus erosivity of melt runoff. For the months with temperatures > 0 °C, the melt runoff erosivity was omitted from the calculations. The R-factor is the product of kinetic energy of rainfall event (E) and its maximum 30-min intensity (I30) (Brown and Foster, 1987, Panagos et al., 2015a):
251.5 mm in July 2011 and the lowest 0.0 mm and 3.5 mm in October 2011 and January 2014, respectively. The highest daily precipitation was 100.7 mm on 15 May 2014. In the years 2011–2014, mean annual air temperature was 6.3 °C, and the hottest year, with mean annual temperature of 7.4 °C, was 2014. The lowest temperature of −25.8 °C was recorded on 3 February 2012 and the highest, 33.0 °C, on 8 July 2013. The snow starts to melt at the beginning of the March. The growing season starts around 10 April. The winter begins around 30 November. The mean number of days with light frost in the catchment area is between 120 and 130. Mean number of days with precipitation is 190, of which 70–80 are snowy days. Soils in the catchment area are diverse, depending on location on the slope. The soil cover in the Mątny stream catchment is dominated by loamy soils, including sandy clay loam, loam, silt loam, clay loam and sandy loam. Pedological conditions were identified by analysis of a 1:25,000 agricultural soil map and qualified in the respective groups according to USDA standards (Soil Survey Staff, 1975).
R=
V = e (LU ∙ Fcover )
Scale 1:5000, resolution 5m Scale 1:25,000
Digital map of soil
Digital map of land use
Meteorological data
– Institute of Meteorology and Water Management (2011–2014)
(5)
where: V — the vegetation retention (dimensionless), with V = 1 for bare land and V > 1 for managed land, LU — an empirical land-use parameter ranging from 1 to 10, and Fcover — a vegetation layer normalized within 0 to 1 range (in accordance with the principles of soil and water conservation). The LU values were derived from the CORINE Land Cover 2012 database (CLC, 2012). There are 33 soil type classes according to CLC, 2012 in Poland; of these, five classes are encountered in the Mątny stream catchment area. Agricultural lands include arable lands outside the reach of water supply systems, meadows and pastures as well as agricultural areas with a high share of natural vegetation. We compared deciduous forests, coniferous forests and mixed forests. Areas with scattered buildings have been excluded from these calculations; such settled area was not a very large factor in this site. The CLC classes were verified on the basis of on-site inspections on 20 July 2011, 25 July 2012, 15 June 2013 and 25 July 2014. The plant cover factor (Fcover) for grassland and forests was established to be 1 (Panagos et al. 2014a). Fcover for crops was established as a cultivated agri-environmental scheme in our study region (based on effective management, vegetation arrangement and individual features of plant growth). In addition, in the case of cereals (spring wheat and spring oats), the stubble field was left unplowed for the winter after the harvest. As regards arable land, potato harvesting was followed by the essential agrotechnical treatments (plowing, harrowing) and white mustard (Sinapis alba) was sown as a cover crop. Thus, the value of Fcover for arable land ranges from 0 to 1, depending on the plant development phase. The values of the plant vegetation cover are shown in Table 2.
Table 1 Categories, sources and description of G2 model for the study area.
Geodesic and Cartographic Documentation Center Institute of Soil Science and Plant Cultivation State Research Institute – Landsat – Orthophotomap – Copernicus Land (CORINE Land Cover 2012 PL) – Site control
(4)
2.3.3. Vegetation retention (V) The vegetation retention factor (V) is the degree to which the vegetation cover and management are expected to protect soil from erosion. In the G2 model, V was calculated based on Panagos et al. (2012, 2014a):
2.3.2. Rainfall erosivity (R) The rainfall-runoff erosivity factor for a specific time period – R –
DEM
−1
where: rain10 is precipitation of ≥10 mm; and day10 — number of days with precipitation above 10 mm.
where: E — the predicted soil loss from an area during a specific time period (Mg ha−1), R — the rainfall-runoff erosivity factor (MJ mm ha−1 h−1 (y/12)−1) for a specific time period which quantifies the impact of raindrop and runoff energy, V — the vegetation retention factor (dimensionless), S — the soil erodibility factor (Mg ha h MJ−1 ha−1 mm−1), T — the topography factor (dimensionless), and I — slope-intercept factor (dimensionless). Calculations for the model algorithm were carried out by means of analyses using ArcGIS 10.3.1 software. The spatial data were based on the Polish State Surveying Coordinate System (PUGW) 1992. Thematic layers and climatic data were developed using the sources specified in Table 1. The DEM basic thematic layer was developed on the basis of point coordinates (X, Y, Z) in a regular network with 1 m resolution, interpolated on the basis of a point cloud using LIDAR aviation laser scanning. DEM with a raster resolution of 5 × 5 m was used (obtained from the Geodesic and Cartographic Documentation Center in Poland).
Data description
(3)
l=1 k=1
EI30 = 7.05∙rain10 − 88.92∙day10
(2)
Source
lm
where: R — the rainfall-runoff erosivity factor (MJ mm ha h (y/ 12)−1), n — number of months covered by the data records, lj — number of erosive events of a given months m, and EI30 — the rainfall erosivity index of a single event k (MJ mm ha−1 h−1). Using this software module, these parameters were calculated for Obidowa station in Poland, for which the precipitation time series of 4 years (2011–2014) took into account precipitation of ≥10 mm (Loureiro and Coutinho, 2001; Suif et al., 2016). The event erosivity EI30 was defined as:
2.3.1. Model description The G2 model formula, expressed as a logical product, combines all the main factors, including natural and anthropogenic, that determine the nature and magnitude of soil erosion. Erosion intensity was assessed from the equation (Panagos et al., 2014a, 2014b):
Data type
n
∑ ∑ (EI30 )k −1
2.3. Model and data
R T E = ⎛ ⎞ ∙S∙ ⎛ ⎞ ⎝V ⎠ ⎝I⎠
1 n
– Multispectral satellite image – Scale 1:10,000 – Resolution 20 m – 20 July 2011 – 25 July 2012 – 15 June 2013 – 25 July 2014 – Weather station in Obidowa
2.3.4. Soil erodibility (S) Soil erodibility, S, was found from the formula proposed by Wischmeier and Smith (1978), modified by Renard et al. (1997) and 118
Catena 164 (2018) 116–124
W. Halecki et al.
image (which is 255 for a 8-bit system). Resolution in a model of the I-factor was between 1.0 and 1.58, with a negligible contribution of the parameter I > 1.42. The average Ifactor for the Mątny catchment was 1.14.
Table 2 The values of the plant vegetation cover. Land use
Spring oat/ common wheat Potatoes
Arable lands
Grasslands Forests
CORINE class
LU
Months
Parameter Fcover (−)
Parameter V (−)
211
6.50
211
6.50
IV V VI VII–III X–V VI VII VIII–IX I–XII I–XII
0.00 0.30 0.60 1.00 0.00 0.10 0.40 0.70 1.00 1.00
1.00 7.02 49.40 665.14 1.00 1,92 13.46 94.63 2980.96 8103.08
231 311/312/313
8.00 9.00
2.3.7. Relative risk index The contribution of the respective months to the annual value of water erosion was assessed from the relative risk index, as presented by Panagos et al. (2014a, 2014b) in the form:
STERIn =
1.14
10−4 ·(12–a) + 3.25·(b–2) + 2.5·(c–3) ⎤ ∙0.1317 ⎥ 100 ⎦ −1
−1
(6) −1
where: S — the soil erodibility factor (Mg ha h MJ ha mm ), M — the texture defined as percentage silt + fine sand fraction content multiplied by (100-clay fraction), a — the organic matter content (%); the a parameter is calculated as: a = 1.72 · orgC, where orgC is the organic carbon content in the soil layer (%), b — the soil texture class (very fine granular: 1, fine granular: 2, medium or coarse granular: 3 and very coarse granular: 4), and c — the profile permeability class (rapid: 1, moderate to rapid: 2, moderate: 3, slow to moderate: 4, slow: 5 and very slow: 6). In soils with a total content of silt and fine sand (grain size 0.002–0.2 mm) above 70%, the value of the soil erodibility factor S was found by means of a nomograph, developed by Wischmeier and Smith (1978). The characteristics required in the model were developed on the basis of soil samples collected from 43 sites in the catchment area. The values of the S factor for the various points ranged from 0.018 to 0.076 Mg ha h MJ−1 ha−1 mm−1. Lower values of the S factor were determined for sandy clay loam (0.018–0.034 Mg ha h MJ−1 ha−1 mm−1) and clay loam (0.022 Mg ha h MJ−1 ha−1 mm−1). For loam, the values were 0.036–0.068 Mg ha h MJ−1 ha−1 mm−1. Values exceeding 0.068 Mg ha h MJ−1 ha−1 mm−1 were recorded for silt soils.
2.3.8. Data processing Redundancy Analysis (RDA) was applied to show a common correlation pattern (Lepš and Šmilauer 2003) for analyzed features from different time periods (Fig. 5). This method might be seen as a promising contribution to validating the use of the main influential factors on water erosion and to document short-term soil stoniness within the model. RDA was used to quantify the importance of factors affecting soil erosion by water. Correlation matrices were selected for land use. The statistical procedure was computed by the program Canoco for Windows ver. 4.51. 3. Results 3.1. Investigation of short-term soil erosion The spatial distribution of yearly erosion risk classes in the investigated area using the G2 model in the respective years from 2011 to 2014 period is shown in a set of maps (Fig. 3). The eroded soil mass (Mg y−1) in the Mątny stream catchment area, as calculated using the G2 model for the respective years in the period 2011–2014 was as follows: 967.92 (2011), 494.87 (2012), 1418.32 (2013), and 4563.87 (2014). Total annual soil loss and water erosion for the years 2011–2014 was 12.7 Mg ha−1. The maximum eroded soil mass was 31.05 Mg per ha of area. The lowest total annual soil loss was 3.37 Mg ha−1, recorded in 2012. The calculated standard deviation for the multi-year period was 12.52 Mg ha−1, and the variability coefficient was 99%. In the year 2011, the lowest value for total soil eroded mass was obtained in October (16.94 Mg). In 2012 the highest value was 238.90 Mg in June. Total soil eroded mass in the other months was similar (March: 20.50 Mg, August: 18.87 Mg, September: 16.26 Mg, October and November: 10.99 Mg). Total soil eroded mass due to water erosion in the summer months (June–August) was 88.13% of the total annual soil loss. In the autumn and winter periods, it accounted for 7.73 and 4.15% of the total annual soil loss, respectively. In the year 2013, the total soil eroded mass share of the months May–December in the total annual soil loss ranged from 0.46% (December) to 57.08% (June). For the year 2014, total soil eroded mass in the spring months (March–May) was 81.62% of the total annual soil loss. In the summer (June–August) and autumn (September–November) the total soil eroded mass was 16.74% and 1.64%, respectively, of total annual soil loss. Data of the relative risk index (STERIn) is presented in Table 3. The relative risk index values for June, July and August were 27.89, 15.09 and 3.48%, respectively. RDA detected that vegetation retention parameter was highly
2.3.5. Topographic influence (T) To estimate the effects of topography on the erosion risk, an equation developed by Moore and Burch (1986) and proposed for use in erosion modeling:
sin β ⎞1.3 A 0.4 LS = T = ⎛ s ⎞ ∙ ⎛ ⎝ 22.13 ⎠ ⎝ 0.0896 ⎠
(7)
where: T — the topographic influence (≥0, dimensionless), As — the flow accumulation (m), and β — the slope (rad). The respective equation parameters were found using the ArcMap 10.3.1 software. The values of the T parameter ranged from 0.00 to 38.79 (the mean value was 4.19). 2.3.6. Slope intercept (I) The slope intercept (I) was estimated by applying a Sobel filter on the high resolution satellite image mosaic. The equation for the computation of the I factor was compared with that proposed by Panagos et al. (2012) in order to include the I-factor in the denominator in the previous equation (Karydas et al., 2015):
I=1+
Sf DNmax
(9)
where: STERIn — relative risk index [–], En — water erosion for the respective months, and E — annual water erosion. The extent of soil erosion was determined as per an erosion risk scale proposed by Marks et al. (1989), which divides erosion risk into six classes: no risk: < 1, very low risk: 1–5; low risk: 5–10, medium risk: 10–15, high risk: 15–30, and very high risk: > 30, expressed in Mg ha−1.
presented by Panagos et al. (2012):
2.1·M S=⎡ ⎢ ⎣
En E
(8)
where: I — the slope intercept [–], Sf — value of digital image mosaic as generated using the Sobel filter with a 3 × 3 mask, adopting values from 0 to DNmax, and DNmax — maximum value of the digital Sobel 119
Catena 164 (2018) 116–124
W. Halecki et al.
Table 3 Total annual soil loss and STERIn in 2011–2014 period. Month/year
January February March April May June July August September October November December
STERIn (%)
Total annual soil loss (Mg) 2011
2012
2013
2014
2011
2012
2013
2014
– – – – 501.25 269.93 145.10 33.71 – 16.94 – –
0.02 20.50 – – – 238.9 178.33 18.87 16.26 10.99 10.99 –
– – – – 477.22 809.57 29.83 14.70 17.20 34.79 28.50 6.52
– – – 56.13 3669.19 606.84 137.06 19.89 41.86 41.86 – –
– – – – 51.79 27.89 15.09 3.48 – 1.75 – –
0.00 4.14 – – – 48.28 36.04 3.81 3.29 2.22 2.22 –
– – – – 33.65 57.08 2.10 1.04 1.21 2.45 2.01 0.46
– – – 1.23 80.40 13.30 3.00 0.43 0.92 0.72 – –
Fig. 2. Distribution of the rainfall erosivity (R factor) for the years 2011–2014.
4. Discussion
correlated with forest, while slope intercept and event erosivity parameters were negatively related with grassland as well as arable area (Fig. 5). The estimated R-factor values (Fig. 2) were characterized by strong seasonality, with the highest regime coefficients occurring from May to July (e.g. May 2013: 110.41 MJ mm ha−1 h−1; May 2014: 838.90 MJ mm ha−1 h−1) and the lowest (almost zero) occurring in March. The mean estimated R-factor for the Mątny catchment was 92.5 MJ mm ha−1 h−1. The mean annual value of the parameter R (MJ mm ha−1 h−1 y−1) for the catchment was 94.68 (2011), 86.34 (2012), 106.02 (2013), and 234.75 (2014).
4.1. Modeling of water erosion at the catchment scale The melt runoff erosivity Rs was found in accordance with the USLE method (Wischmeier and Smith, 1978), assuming that its value for the first month in a year with temperatures > 0 °C (March) is 1/10 of the total precipitation layer, as expressed in millimeters of water column in the preceding period of time with temperatures < 0 °C. The calculated values of Rs were 5.6 mm (2011), 17.2 mm (2012), 16.2 mm (2013), and 4.2 mm (2014). In Poland, the method for the Rs factor assessment was also used e.g. by Banasik and Górski (1990), and Kruk et al. (2016). The monthly distribution of rainfall-runoff erosivity factor R for a specific time period is shown in Fig. 2. The value of the R factor seems closely correlated with total annual precipitation. The relationship is described by the power function, fitted with the correlation coefficient r = 0.76. The effect of kinetic energy of rainfall EI30 on the value of the R factor was characterized by the correlation coefficient r = 0.91. Similar seasonal patterns were reported by Świętochowicz (2013) for the Limanowa station in Beskid Wyspowy (amounting to 113.5 (MJ mm ha−1 h−1 y−1) and a model fit (correlation coefficient — R) equal to 0.75) where winter plowing was used. According to the ESDAC program (European Soil Data Centre), which is based on the RUSLE2015 program and the EIONET-SOIL project (Joint Research Centre, European Commission), the Mątny stream catchment area generates a total soil eroded mass of 5–10 Mg ha−1 y−1. Kowalczyk and Twardy (2012) indicated the role of the soil and water conservation and the effects of inappropriate management in the
3.2. Erosion intensity Erosion intensity distribution in the respective months was analyzed against the land-use (Fig. 4). The highest total annual soil losses were recorded in 2014. For spring oats, the losses in 2014 were 9.27 Mg ha−1, for potatoes 650.78 Mg ha−1, for grassland 3.24 Mg ha−1, and for forests 0.06 Mg ha−1. With reference to land cover, the highest soil losses were recorded in the spring and summer months on arable lands where root plants were cultivated. The values ranged from 68.31% (2012) to 90.12% (2014) of the total annual soil loss. The contribution of grassland to the annual water erosion ranged from 31.11% (2012) to 7.68% (2014) of the total annual soil loss. For the years in this study, the total soil eroded mass in forest covered areas was negligible, not higher than 0.1% of its total value.
120
Catena 164 (2018) 116–124
W. Halecki et al.
Fig. 3. Variability of soil loss (detachment soil participles through transport and surface-runoff) caused by water erosion in consecutive years.
catchment of the Bystrzyca Dusznicka (197.7 km2) in the south-west of Poland. In this study a mean total annual soil loss was 900.35 Mg km−2. Panagos et al. (2014a) studied soil erosion using the G2 method on Crete with precipitation data from 24 weather stations for the multiyear period 1969–1979. Water erosion assessment led to soil loss underestimation, especially for croplands and grasslands. It is, therefore, still of prime importance to improve the newly developed method. In our erosion model, we integrated all factors connected with total soil eroded mass. In further research, the G2 model can be used to predict soil erosion even under natural vegetation. Karydas et al. (2015) performed complex studies on actual soil erosion using the G2 model for various types of crops. Mean monthly values of potential erosion intensity are essential for appraisal of the potential risk of water erosion. This is one of the principal advantages of the G2 model. Our study showed that soil loss assessment is substantial to control erosion by water. We suggest that soil and water conservation is key in reducing soil erosion in arable areas. However, erosion
intensity can be increased in the crops such as potatoes, and nearby native evergreen forests in erosion-prone area. 4.2. Consequences of land-use changes for soil erosion by water It is crucial to design appropriate efficient tools for water erosion and land management (Bosco et al., 2015; Jha et al., 2015). Obviously, predicting runoff distribution has been presented in the literature where soil erosion intensively occurs and clearly points towards the importance of empirical and theoretical models (White et al., 2011; Tilahun et al., 2014; Rathjens et al., 2015; Hoang et al., 2017). Notoriously, in the European Union countries, mean values were very close to the 2.46 t ha1 yr1 analytically estimated in erosion-prone lands. In our study, mean annual soil erosion was between 3.37 and 31.04 t ha1 yr1. To mitigate land degradation processes, there is a need to provide the reduction of soil erosion by water based on support practices — P-factor (Panagos et al., 2015a,b). Moreover, spatial 121
Catena 164 (2018) 116–124
W. Halecki et al.
Fig. 4. Temporal trends (monthly mean) of R (rainfall erosivity), V (vegetation retention) and E (erosion intensity) parameters for the years 2011–2014.
Table 4 R-value according to our method (minimum, maximum and mean in the period 2011–2014) and Ballabio et al. (2017). Month
Minimum R
Maximum R
Mean R 2011–2014
Ballabio et al. (2017)
January February March April May June July August September October November December
0 0 0 0 110.41 103.89 61.16 81.6 0 52.08 0 0
0.1 97.2 0 56.92 848.97 352.05 365.63 145.76 181.02 164.88 135.09 30.93
0 24 (+3↑) 0 14 269 (+168↑) 209 252 94 (−62↓) 81 112 47 (+14↑) 8
– < 21 < 21 < 21 66–99 > 156 > 157 > 156 66–99 99–158 21–33 < 21
↑ means overestimated value; ↓ illustrate underestimated value.
analysis of soil features is helpful in determination of factors affecting the saturation-excess runoff type in mountain areas, particularly for sustainable agricultural practices and land use (Halecki et al., 2016; Halecki et al., 2017). In Polish conditions water erosion is highly dependent on recent vegetative phase. In a case of oat and wheat the value of V parameter attains the minimum on April, and the maximum occurs between July and March. This tendency is the same for potatoes, however, in another periods of total soil eroded mass. Monthly erosive periods are not influential on grasslands and forests (Table 4). Our study revealed that land-cover corresponded with soil erodibility, rainfall erosivity and vegetation retention in a profound way (Fig. 5). This is highly probable because, nowadays, not all farmers tend to implement erosion-limiting treatments, such as leaving unplowed stubble field over the winter, or the use of cover crops.
Fig. 5. RDA plot. Solid squares represent arable area, rectangles indicate grassland and diamonds designate — forest zone. Eigenvectors (arrows) stand for factors affecting soil erosion by water in different land use. The two components of the RDA analysis explained 72.8% of a total variability. Explained variance for the first axis was 40.8% and for the second one was 32.0%.
for the first time under conditions prevalent in Poland. Undoubtedly the advantage of the model is the possibility of assessing total soil eroded mass over monthly intervals. It enables the determination of the influence of plant growing stages on the V parameter shaping. In the work, details from the CLC2012 database, based on satellite orthophotomaps on field observation, was demonstrated. A limitation on the model is the resolution of maps from CORINE Land Cover database. The G2 model introduces a new quality in the assessment of soil erosion intensity. The monthly approach of the model using various plant stages was important for soil and water conservation. The study verified the stochastic method of estimating soil erosion under different regimes of sediment transport and land use. The presented results indicated the relative consequence of land use. The G2 model, as we conjecture from our preliminary investigations, seems to be dominated by the creation of short-term erosive events and the detachment of soil material. Our calculations have also showed the important role of estimating conservation measures against water erosion, and focused on soil erodibility. We suggest that the G2 model is a useful tool for surface water erosion appraisal such as rain drop and rill erosion on small-scale in short time intervals. Long-term data series should be implemented in
5. Conclusion The monthly approach to the G2 model enabled an identification of the months that generated the highest total soil eroded mass in Polish mountainous weather conditions. Redundancy analysis based on landuse types displayed factors affecting soil erosion by water. The highest total soil eroded mass was recorded in the spring and summer months on arable land planted with potatoes. The annual value of water erosion in the catchment was also found to be negatively affected by the presence of grassland, in addition to cultivated plants. This is associated with sheep- and cattle grazing in the area. In contrast, total soil eroded mass was negligible in the forests. Here, the G2 model has been verified 122
Catena 164 (2018) 116–124
W. Halecki et al.
further investigations to improve modeling and calibration schemes by simulating high rates of soil erosion (rill erosion) on a larger scale.
GIS in a Shivalik subwatershed. J. Earth Syst. Sci. 122 (2), 389–398. Laflen, J.M., Moldenhauer, W.C., 2003. Pioneering soil erosion prediction: the USLE story. In: Word Association of Soil and Water Conservation Special Publication. 1. pp. 54. Lee, G.S., Choi, I.H., 2010. Scaling effect for the quantification of soil loss using GIS spatial analysis. KSCE J. Civ. Eng. 14 (6), 897–904. http://dx.doi.org/10.1007/ s12205-010-0878-x. Lepš, J., Šmilauer, P., 2003. Multivariate Analysis of Ecological Data Using CANOCO. Cambridge University Press, New York. Loureiro, N.S., Coutinho, M.A., 2001. A new procedure to estimate the RUSLE EI30 index, based on monthly rainfall data and applied to the Algarve Region Portugal. J. Hydrol. 250, 12–18. Mao, D., Cherkauer, K.A., 2009. Impact of land-use change on hydrologic responses in the Great Lakes region. J. Hydrol. 374, 71–82. http://dx.doi.org/10.1016/j.jhydrol. 2009.06.016. Marks, R., Müller, M.J., Leser, H., Klink, H.J., 1989. Anleitung zur Bewertung des Leistungsvermögens des Landschaftshaushaltes (BA LVL). In: Zentralausschuß für deutsche Landeskunde. Forschungen zur Deutschen Landeskunde. 229. Mhangara, P., Kakembo, V., Lim, K., 2012. Soil erosion risk assessment of the Keiskamma catchment, South Africa using GIS and remote sensing. Environ. Earth Sci. 65 (7), 2087–2102. Moore, I.D., Burch, G.J., 1986. Sediment transport catchment of sheet and rill flow: application of unit stream power theory. Water Resour. Res. 27 (8), 1350–1360. Nadal-Romero, E., Regüés, D., Latron, J., 2008. Relationships among rainfall, runoff, and suspended sediment in a small catchment. Catena 74 (2), 127–136. http://dx.doi. org/10.1016/j.catena.2008.03.014. Panagos, P., Karydas, C.G., Gitas, I.Z., Montanarella, L., 2012. Monthly soil erosion monitoring based on remotely sensed biophysical parameters: a case study in Strymonas river basin towards a functional pan-European service. Int. J. Digital Earth 5 (6), 461–487. https://doi.org/10.1080/17538947.2011.587897. Panagos, P., Karydas, C.G., Ballabio, C., Gitas, I.Z., 2014a. Seasonal monitoring of soil erosion at regional scale: an application of the G2 model in Crete focusing on agricultural land uses. Int. J. Appl. Earth Obs. Geoinf. 27PB, 147–155. http://dx.doi.org/ 10.1016/j.jag.2013.09.012. Panagos, P., Meusburger, K., Ballabio, C., Borrelli, P., Alewell, C., 2014b. Soil erodibility in Europe: a high-resolution dataset based on LUCAS. Sci. Total Environ. 479 –480, 189–200. http://dx.doi.org/10.1016/j.scitotenv.2014.02.010. Panagos, P., Ballabio, C., Borrelli, P., Meusburger, K., Klik, A., Rousseva, S., Tadić, M.P., Michhaelides, S., Hrabalíková, M., Olsen, P., Aalto, J., Lakatos, M., Rymszewicz, A., Dumitrescu, A., Beguería, S., Alewell, C., 2015a. Rainfall erosivity in Europe. Sci. Total Environ. 511, 801–814. http://dx.doi.org/10.1016/j.scitotenv.2015.01.008. Panagos, P., Borrelli, P., Poesen, J., Ballabio, C., Lugato, E., Meusburger, K., Montanarella, L., Alewell, C., 2015b. The new assessment of soil loss by water erosion in Europe. Environ. Sci. Pol. 54, 438–444. Panagos, P., Borrelli, P., Meusburger, K., van der Zanden, E.H., Poesen, J., Alewell, C., 2015c. Modelling the effect of support practices (P-factor) on the reduction of soil erosion by water at European Scale. Environ. Sci. Pol. 51, 23–34. Park, S., Oh, C., Jeon, S., Jung, H., Choi, C., 2011. Soil erosion risk in Korean watersheds, assessed using the revised universal soil erosion equation. J. Hydrol. 399, 263–273. http://dx.doi.org/10.1016/j.hydrol.2011.01.004. Perović, V., Životić, L., Kadović, R., Đorđević, A., Jaramaz, D., Mrvić, V., Todorovic, M., 2013. Spatial modelling of soil erosion potential in a mountainous watershed of South-eastern Serbia. Environ. Earth Sci. 68, 115–128. Podwojewski, P., Orange, D., Jouqert, P., Valentin, C., Nguyen, C.T., 2008. Land-use impacts on surface runoff and soil detachment within agricultural sloping lands in Northern Vietnam. Catena 74, 109–118. http://dx.doi.org/10.1016/j.catena.2008. 03.013. Rathjens, H., Oppelt, N., Bosch, D.D., Arnold, J.G., Volk, M., 2015. Development of a gridbased version of the SWAT landscape model. Hydrol. Process. 29, 900–914. Renard, K.G., Foster, G.R., Weesies, G.A., McCool, D.K., Yoder, D.C., 1997. Predicting soil erosion by water: a guide to conservation planning with the Revised Universal Soil Loss Equation (RUSLE). In: Agriculture Handbook. 73. US Department of Agriculture, Washington, DC, pp. 1–251. Smith, S.J., Wiliams, J.R., Menzel, R.G., Coleman, G.A., 1984. Prediction of sediment yield from southern plains grasslands with the Modified Universal Soil Loss Equation. J. Range Manag. 37, 295–297. Soil Survey Staff, 1975. Soil Taxonomy. USDA, Washington, pp. 17–754. Suif, Z., Fleifle, A., Yishimura, C., Saavedra, O., 2016. Spatio-temporal patterns of soil erosion and suspended sediment dynamics in the Mekong River Basin. Sci. Total Environ. 568, 933–945. Świętochowicz, J., 2013. Mean annual R factor for Łazy near Bochnia meteorological station (Carpathian Foothills climatic region). Landf. Anal. 24, 85–95. Tilahun, S.A., Guzman, C.D., Zegeye, A.D., Ayana, E.K., Collick, A.S., Yitaferu, B., Steenhuis, T.S., 2014. Spatial and Temporal Patterns of Soil Erosion in the Semihumid Ethiopian Highlands: A Case Study of Debre Mawi Watershed. Springer, Nile River Basin, pp. 149–163. Ustun, B., 2008. Soil erosion modelling by using GIS and remote sensing: a case study Ganos Mountain. Int. Arch. Photogramm. Remote. Sens. Spat. Inf. Sci. 37, 681–1684. Verstraete, G., Poesen, J., 2001. Factors controlling sediment yield from small intensively cultivated catchments in a temperate humid climate. Geomorphology 40 (1), 123–144. http://dx.doi.org/10.1016/S0169-555X(01)00040-X. White, E.D., Easton, Z.M., Fuka, D.R., Collick, A.S., Adgo, E., McCartney, M., Steenhuis, T.S., 2011. Development and application of a physically based landscape water balance in the SWAT model. Hydrol. Process. 25, 915–925. http://dx.doi.org/10. 1002/hyp.7876. Wischmeier, W.H., Smith, D.D., 1978. Predicting rainfall erosion losses — a guide to
Acknowledgements This research was financed by the Ministry of Science and Higher Education of the Republic of Poland. The study was also supported by the Projects No. DS-3331/KMiKS/2011-2017. References Bagarello, V., Ferro, V., Pampalone, V., 2015. A new version of the USLE-MM for predicting bare plot soil loss at the Sparacia (South Italy) experimental site. Hydrol. Process. 29, 420–4219. http://dx.doi.org/10.1002/hyp.10486. Bahadur, K.C.K., 2009. Mapping soil erosion susceptibility using remote sensing and GIS: a case of the Upper Nam Wa Watershed, Nan Province, Thailand. Environ. Geol. 57, 695–705. http://dx.doi.org/10.1007/s00254-008-1348-3. Ballabio, C., Borrelli, P., Spinoni, J., Meusburger, K., Michaelides, S., Beguería, S., Klik, A., Petan, S., Janeček, M., Olsen, P., Aalto, J., Lakatos, M., Rymszewicz, A., Dumitrescu, A., Tadic, M.P., Diodato, N., Kostalova, J., Rousseva, S., Banasik, K., Alewell, C., Panagos, P., 2017. Mapping monthly rainfall erosivity in Europe. Sci. Total Environ. 579, s.1298–1315. Banasik, K., Górski, D., 1990. Wyznaczanie erozyjności deszczy do uniwersalnego równania strat glebowych. In: Zeszyty Naukowe AR Wrocław, Melioracje. 34. pp. 103–109. Bosco, C., de Rigo, D., Dewitte, O., Poesen, J., Panagos, P., 2015. Modelling soil erosion at European scale: towards harmonization and reproducibility. Nat. Hazards Earth Syst. Sci. 15, 225–245. Brown, L.C., Foster, G.R., 1987. Storm erosivity using idealized intensity distributions. Trans. ASAE 30, 379–386. Burrough, P.A., 2001. GIS and geostatistics: essential partners for spatial analysis. Environ. Ecol. Stat. 8, 361–377. http://dx.doi.org/10.1023/A:1012734519752. Cârdei, P., 2010. The dimensional analysis of the USLE — MUSLE soil erosion model. Proc. Rom. Acad., Ser. B 3, 249–253. CLC, 2012. CORINE Land Cover. http://inspire.gios.gov.pl/portal/index.php?profile= 29542&projection=EPSG%3A2180. Dabral, P.P., Baithuri, A., Pandey, A., 2008. Soil erosion assessment in a hilly catchment of North Eastern India using USLE, GIS and remote sensing. Water Resour. Manag. 22, 1783–1798. http://dx.doi.org/10.1007/s11269-008-9253-9. Flanagan, D.C., Lafen, J.M., Meyer, L.D., 2003. American Society of Agricultural Engineers Historic Landmark Dedication; Honoring the Universal Soil Loss Equation. Agric. & Biol. Eng., Purdue University. Halecki, W., Młyński, D., Ryczek, M., Kruk, E., Lis, J., 2016. Zmienność przestrzenna wilgotności i gęstości objętościowej gleby w zlewni potoku górskiego Mątny w Gorcach. In: EPISTEME. 30. pp. 347–358 ISSN 1895-4421. Halecki, W., Młyński, D., Ryczek, M., Kruk, E., Radecki-Pawlik, A., 2017. The application of Artificial Neural Network (ANN) to assessment of soil salinity and temperature variability in agricultural areas of a mountain catchment. Pol. J. Environ. Stud. 6, 2545–2554. Hoang, L., Schneiderman, E.M., Moore, K.E.B., Mukundan, R., Owens, E.M., Steenhuis, T.S., 2017. Predicting saturation-excess runoff distribution with a lumped hillslope model: SWAT-HS. Hydrol. Process. 31, 2226–2243. http://dx.doi.org/10.1002/hyp. 11179. Jha, A., Schkade, U., Kirchner, G., 2015. Estimating short term soil erosion rates after single and multiple rainfall events by modelling the vertical distribution of cosmogenic 7Be in soils. Geoderma 243–244, 149–156. Karydas, C.G., Panagos, P., 2016. Modelling monthly soil losses and sediment yields in Cyprus. Int. J. Digital Earth 9 (8), 766–778. Karydas, C.G., Panagos, P., 2018. The G2 erosion model: an algorithm for month-time step assessments. Environ. Res. 161, 256–267. http://dx.doi.org/10.1016/j.envres. 2017.11.010. Karydas, C.G., Panagos, P., Gitas, I.Z., 2014. A classification of water erosion models according to their geospatial characteristics. Int. J. Digital Earth 7 (3), 229–250. http://dx.doi.org/10.1080/17538947.2012.671380. Karydas, C.G., Zdruli, P., Koci, S., Sallaku, F., 2015. Monthly time-step erosion risk monitoring of Ishmi-Erzeni watershed, Albania, using the G2 model. Environ. Model. Assess. 20 (6), 657–672. http://dx.doi.org/10.1007/s10666-015-9455-5. Kinnell, P.I.A., 2016. Comparison between the USLE, the USLE-M and replicate plots to model rainfall erosion on bare fallow area. Catena 145, 39–464b. http://dx.doi.org/ 10.1016/j.catena.2016.05.017. Kinnell, P.I.A., Risse, L.M., 1998. USLE-M: empirical modeling rainfall erosion through runoff and sediment concentration. Soil Sci. Soc. Am. J. 62, 1667–1672. Kondracki, J., 2000. Geografia regionalna Polski. PWN, Warszawa. Kowalczyk, A., Twardy, S., 2012. Wielkość erozji wodnej obliczona metodą USLE. WodaŚrodowisko-Obszary Wiejskie 12 (37), 83–92. Kruk, E., 2017. Influence of daily precipitation on yield of eroded soil in mountain basin using the Musle model. Acta. Sci. Pol. Form. Circum. 16 (2), 147–158. http://dx.doi. org/10.15576/ASP.FC/2017.16.2.147. Kruk, E., Ryczek, M., Malec, M., 2016. Porównanie wielkości strat glebowych wyznaczonych modelami USLE i RUSLE w zlewni górskiej. Wybrane aspekty badawcze. In: Aspekty teoretyczno-praktyczne badań naukowych. 3. pp. 55–65. Kumar, S., Kushwaha, S.P.S., 2013. Modeling soil erosion risk based on RUSLE-3D using
123
Catena 164 (2018) 116–124
W. Halecki et al.
Korçe region, Albania. Earth Sci. Inf. 9 (4), 481–495. Zhang, Y., Degroote, J., Wolter, C., Sugumaran, R., 2009. Integration of modified universal soil loss equation (MUSLE) into a gis framework to asses soil erosion risk. Land Degrad. Dev. 20, 84–91. http://dx.doi.org/10.1002/ldr.893. Żmuda, R., Sasik, J., Szewrański, S., 2005. Analiza potrzeb zmian zagospodarowania przestrzennego wzgórz trzebnickich w aspekcie ochrony przed erozją wodną gleb. Acta Agrophysica 5 (1), 229–237.
conservation planning. In: Supersedes Agriculture Handbook No. 282, pp. 4–11 Washington. Zabaleda, A., Martinez, M., Mriarte, J.A., Antiguedad, I., 2007. Factors controlling suspended sediment yield during runoff events in small headwater catchments of the Basque country. Catena 71 (1), 179–190. http://dx.doi.org/10.1016/j.catena.2006. 06.007. Zdruli, P., Karydas, C.G., Dedaj, K., Salillari, I., Cela, F., Lushaj, S., Panagos, P., 2016. High resolution spatiotemporal analysis of erosion risk per land cover category in
124