Potential Soil Erosion Simulation On Tenerife With Use Of The Revised Universal Soil Loss Equation

(1)

June

2015

Potential Soil Erosion

Simulation On Tenerife With

Use Of The Revised Universal

Soil Loss Equation

Lorenzo Turk

S u p e r v i s o r s : d h r . d r . K e n n e t h R i j s d i j k & d h r . d r . E r i k C a m m e r a a t

(2)

Abstract:

Soil erosion causes loss of organic rich topsoils. It can cause severe financial and irreversible damage to (agricultural) soils. It is therefore important to know which soils are prone to soil erosion. On Tenerife, approximately 42% of the soil surface is affected by rainfall accelerated soil erosion. Tenerife is the biggest of the 7 Canary Islands. It has a great climatic variety with a humid northern and an arid southern part. For this research, the potential soil erosion on Tenerife is simulated with use of the revised universal soil loss equation (RUSLE). The RUSLE equation is implemented in a Geographic Information System (GIS) with use of Arcgis software. The parameter values are calculated using equations and literature. Two scenarios are used in the simulation: the current situation and an average global temperature rise of 2°C situation. Besides, a sensitivity analysis was conducted in order to analyze the effect of input value ranges. The results exposed 10 erosion hotspots on Tenerife; five of them are big gullies and the other five different natural areas. The future scenario simulation resulted in a slight decreasing, but not significant, change in predicted potential soil erosion. The results of the sensitivity analysis variated. The soil parameter showed more stable results than the land cover parameter.

(3)

Relevance

Trimble’s (1975) definition of soil erosion is ‘the total amount of soil material dislocated and removed some distance by erosion within an area’. Soil erosion is harmful because of loss of organic rich topsoils. Especially on agricultural soils the financial damage can be severe because of this process. Besides, when losing more than 1 t/ha/year the effect is irreversible in the next 50-100 years (Van der Knijff et al., 2000). On Tenerife, approximately 42% of the soil surface was affected by rainfall accelerated soil erosion at the beginning of this century (Rodríguez Rodríguez et al., 1998). According to the Tenerife Island Council, agriculture is an important factor for the island because it provides sustainable and cultural benefits; despite it only contributes for 10% of the GDP (“Agricultual Sector”, n.d.). It is therefore important to know which spots on Tenerife are vulnerable to soil erosion.

This research is conducted in order to get a clear overview of the vulnerable spots on Tenerife regarding potential soil erosion during the year. Besides, the expected future rainfall scenario gives an indication of potential change of soil erosion vulnerability. Furthermore, this research is compared with soil loss research in Mauritius from Norder (2010), in order to determine whether the patterns of the results from that simulation are specific for Mauritius or generic.

Introduction

In the introduction section the climate, geology, soils and land use of Tenerife are described. After that the relevance, goals and research questions are discussed.

Tenerife

Tenerife is with 2058 km2_{the largest of the seven big Canary Islands. It lies} between 27° and 30° N latitude and 13° and 18° longitude, west of the border between Morocco and Western Sahara, as showed in figure 1. Currently, around 900.000 people live on the island. Tenerife has the largest agricultural production of the Canary Islands (Diaz-Diaz et al., 1999).

(5)

Climate

Tenerife has a large climatic variety. Northeastern winds bring wet air from the Atlantic Ocean, dry air is coming from the Sahara. The wet air cannot reach the southern side of the island because of the altitude of volcanoes (up to 3717m). As a result, the northern side of the island humid (625mm annual rainfall) and the south side of the island is arid (69 m annual rainfall) (Guera et al., 2006).

Because of the uneven distribution of rain over the year as shown if figure 2, there are no permanent waterways on Tenerife (Casalí, & Giménez, 2007). The rain falling in the wet months is transported via gullies (Spanish: barrancos). Geology

The island is built up from volcanic rocks, except from some local sedimentary horizons. According to Borley (1974), the volcanicity can be split up in three time periods. The early vulcanicty (15.2Ma – 7.2Ma) introduce the existence of the island with shallow slopes and low chemical differentiation in the volcanic rocks. After that, the ‘Vilaflor Complex’ was constructed. During this period, much more

5

Figure 1: A map of the location of the Canary Islands and Tenerife (De Nascimento et al., 2009).

Figure 2: Spatial and annual rainfall distribution on Tenerife (Herrera et al., 2001)

(6)

explosive eruptions took place. Estimations of the maximum height of the complex varies between 3000m to 5000m (Booth, 1973). During the recent series of (explosive) volcanic eruptions (2.58Ma – present) the Teide (3717m) and Viejo (3303m) volcanoes emerged. In historic times, six small eruptions occurred (Lagunilla & Pacheo, 1987).

Soils

The great variety of soils can be clarified by the great differences in altitude and climate. The following main soil types of Tenerife were derived from a soil map from the UvA (2005): andosols, anthrosols, arenosols, cambisols, fluvisols, leptosols, luvisols, regosols, solonetz and vertisols. Immature soils such as arenosols, leptosols and regosols can mainly be found on steep slopes on the southern part of the island, but they are also present in the northern part. Fertile soils such as andosols, cambisols, fluvisols, luvisols, and the soils that are strong influenced by agriculture, anthrosols, can mainly be found at the northern side of the island. They are also present in the south; especially cambisols, which are developed fertile soils.

Land Use

Apart from the urban areas, many natural areas can be found on Tenerife. Multiple types of forests are present (broad-leaved, deciduous, conifer), just as areas with woodland and scrub, moors and heathland and sclerophyllous vegetation. The latter is a type of vegetation with hard leaves that are resistant to drought. As far as agriculture is concerned, bananas, tomatoes, potatoes, vines and orchards are cultivated.

Goals and Research Questions

This research is executed to determine the potential soil erosion hotspots on Tenerife with use of the revised universal soil loss equation (RUSLE) equation in a geographic information system (GIS), and to explore future change in potential soil erosion with respect to climate change. The sensitivity of the model with respect to the range of values for the crop management and vegetation factor and the soil erodibility factor will be examined.

In order to fulfill these goals, the following main research question have been formulated:

 What is the magnitude of potential erosion and where are the hotspots located?

To be able to answer the main research question properly, the following sub questions will also be answered first:

 What factors influence the locations of hotspots?

 To what extend do uncertainty ranges of the parameters influence the results?

 How can a future situation simulation be compared to the current situation simulation?

(7)

Method

Introduction RUSLE

To simulate potential soil erosion the revised universal soil loss equation (RUSLE) from Renard et al. (1991) is implemented in a geographic information system (GIS) with Arcgis 10.1. The RUSLE “predicts the long-term average annual rate of erosion on a field based on soil erosion factors” (Kouli, Soupios & Vallianatos, 2009). Although the RUSLE equation was originally developed for soils in the USA, the equation is widely used in other areas, such as Mauritius (Norder, 2010), Greece (Kouli, Soupios & Vallianatos, 2009), Kenya (Angima et al., 2003) and Philippines (Alejandro & Omasa, 2007). The RUSLE equation consists of the following parameters as described:

A = R * K * LS * C * P (1)

Where

A: The total potential soil loss [t ha-1_year-1_]

R: The rainfall-runoff erosivity factor [MJ mm ha-1_h-1_year-1_{]. This is the impact of} rainfall that causes soil erosion. This will be derived from average amounts of rainfall for each month.

K: The sensitivity of erosion for specific soil-types [t ha MJ-1_mm-1_].

LS: The effect of topography on soil erosion follows from the LS factor. It consists of a slope-length (L) and a soil-steepness (S) parameter.

C: The vegetation cover and management factor C represents the effect of cropping and management practices in agricultural management, and the effect of ground, tree, and grass covers on reducing soil loss in non-agricultural situations.

P: A supporting practices factor. This is the ratio between soil loss with support practices like strip-cropping or contouring, and the soil loss of straight-row farming up and down the slope (Kamminga, 2008). This factor is often omitted from the calculation due to absence of relevant data.

The GIS-model that is used for the simulation is from Kamminga (2008). A screenshot of the model can be found in appendix 1. This this was originally developed for Mauritius, therefore it had to be adapted for Tenerife. The calculation of the parameters and the implementation into the GIS-model will be described in the next paragraphs.

Rainfall Factor (R)

Preferably, the rainfall intensity is used to calculate the rainfall factor. Wischmeier & Smith (1978) constructed the EI parameter, where E is the total storm energy and I the maximum 30-min intensity. Unfortunately, there is often no information about storm intensity. A good approximation can be made with the monthly and annually amount of rainfall. This method is widely used in RUSLE simulations (Angima et al., 2003; Bagherzadeh, 2014; Kouli et al., 2009).

R = r(F)a ₍₂₎

Where

R is the rainfall factor

r and a location specific values

(8)

The F-index was developed by Fournier (1960) and gives a good approximation of rainfall induced soil erosion (Ferro et al., 1991).

F=

∑

pi

2

P

(3)

Where

pi = the amount of rainfall for a specific month in [mm] p = the yearly amount of rainfall in [mm]

The monthly and annual rainfall maps are downloaded from the Clima Impacto, a working group from the Canary Islands for sustainable development. The input data for the maps originates from weather stations on Tenerife and are measurements from the period 1981-2010. The data is corrected to prevent occurring of negative values when the data is interpolated.

The maps are originally contour maps with precipitation values in mm. To prepare the maps for the model, the contour maps with rainfall amounts had to be interpolated and converted to raster maps. This is done by the ‘topo to raster tool’. Despite the statement from Clima Impacto that the data is corrected to prevent negative values, the interpolated maps consisted of some low negative values. These values were set to zero to prevent failing calculations. Because of the amount of cells and the small negative values, it was assumed that setting them to zero would not give significant other results.

Soil Erodibility Factor (K)

Wischmeier & Smith (1978) constructed a nomograph that is used to calculate the soil erodibility:

K = 2.1*10-4_M1.14_{(12 - a) + 3.25 (b - 2) + 2.5 (c - 3)} ₍₄₎ Where

M = the particle-size parameter, which equals percent silt (0.1-0.002 mm) times the quantity 100-minus-percent-clay.

a = percent organic matter

b = the soil-structure code used in soil classification c = the profile permeability class

This equation is used when empirical data is available. Unfortunately, there was no empirical data available. Therefore, the values had to be approximated with use of soil information from the World Reference Base for soil resources (IUSS Working Group, 2014) and literature. The values are estimated on texture properties of different soils. This is not the most accurate way to estimate erodibility values but sufficient enough for the purpose of this thesis. Some soil types have an explicit soil texture with a small variability in erodibility values. For instance, vertisols have a clay to heavy clay texture. The erodiblity values for these soils range from 0.21 to 0.24, which is an acceptable range. These values were compared to values from different literature to check whether the

(9)

estimation was in the same range. In absence of empirical data, this is the best approximation possible.

For other soil types with a less distinct texture class, literature gave more reliable results than the approximation technique. Most of the units of the soil map that is used are soil complexes; a mix of multiple soil types. For the calculation of the values of these complexes, the values of the main soil types were averaged. The values of all soil complexes can be found in appendix 2.

Implementing the soil map in the GIS-model was done by first assigning the soil types with an erosion value. The map had to be converted into a raster before implementing it into the model. The map was

downloaded via UvA geoportal. It is a digitized version of a

by students in 1982 created soil map (“Soil map of Tenerife”, n.d.).

Topographic Factor (LS)

The topographic factor LS can be divided in two separated factors: the slope length L-factor and slope steepness S-factor. The formula to calculate the L-factor is the following:

L = (λ/22.1)M ₍₅₎

Where

L is the slope length factor λ is the slope length [m]

M is depended on the steepness of the slopes. The assigned values can be found in table 1.

The S-factor is calculated with the following formulas:

S = 10sinθ+0.03 for θ < 9° (6.1)

S = 16.8sinθ-0.50 for θ > 9° (6.2)

Where

θ is the slope angle

The Digital Elevation model is downloaded from the UvA Geoportal and is from the Centro Nacional de Información Geográfica.

In the GIS-model, the slope length and slope steepness are calculated separately. A slope ends at the point where water enters a channel that is part of a drainage system or the steepness changes in such degree, that water stagnates (Kamminga, 2008). After using fill tool to increase the values of extreme low-valued cells and flow direction to decide which direction the erosion occurs, the flow length was calculated.

At the calculation of slope steepness, slopes above and under 9° are distinguished. These are calculated separately and after that, merged to one map. After merging, the maps for the slope length and slope steepness are multiplied to create the topographic factor map.

9

Table 1: Values for M and the corresponding slope levels Slope Value >5° 0.5 3.5-4.5° 0.4 1-3° 0.3 <1 0.2

(10)

Land Cover and Crop Management Factor (C)

The values for the land cover and crop management factor are collected from literature. The land cover map that is used for this factor is from Centro Nacional de Información Geográfica (2006) following the Corine land use classification. Implementing the land cover factor in GIS is done in the same way as the soil erodibility factor. The values were assigned to the land use units. The values can be found in appendix 2. Subsequently, the map could be implemented into the model. After preparing all the factors, the maps were multiplied.

Future Prospect

The future prospect scenario is based on what is called scenario D by Clima Impacto, a global average temperature rise of 2°C. The latest IPCC report states that it is likely that the temperature rises 2°C between now and 2100 (Stocker et al., 2013). In general, when the temperature rises, the amount of rainfall will rise due to enhanced evapotranspiration. All other parameters are set equal to the current situation simulation. This scenario is executed to see if and to what extend the potential erosion will change in the future.

(11)

Results

In this section, the results of the parameter calculation will be evaluated at first. Secondly, the modeling results will be discussed with use of the predicted potential soil erosion maps. Here, the predicted potential erosion hotspots and the potential erosion values will be reviewed. The implementation of the results can be found in the discussion section.

Parameter Results

In table 2 the results of the calculation of the rainfall factor can be found. The monthly maxima, means, standard deviations and the percentages of the cumulative mean can be found. As expected, the rainfall factor is close to zero for almost the whole island from April until September. These six months contribute for 0,45% of the total rainfall factor. It is therefore interesting to have a look at the other months and especially January and December, contributing together for 74.96% to the total rainfall factor. The standard deviation is, except for January, bigger than the difference between the mean and zero. This indicates that the main part of the island has a rainfall factor value between zero and mean, but that there are extreme outliers. The maximum values confirm this statement. Month Maximum [MJ mm ha-1_h-1_year -1_] Mean [MJ mm ha-1_h-1_year -1_] Standar d deviati on % of cumulat ive mean January 33.5606 4.0644 2.4050 35.72 February 20.1257 0.6822 1.3598 5.99 March 11.1840 0.6275 0.9805 5.52 April 0.8511 0.0458 0.0812 0.40 May 0.1288 0.0037 0.0110 0.03 June 0.0796 0.0009 0.0032 0.01 July 0.0027 0.0001 0.0003 0.00 August 0.0002 0.0000 0.0000 0.00 Septemb er 0.0511 0.0016 0.0038 0.01 October 3.6264 0.2649 0.4151 2.33 Novembe r 21.8803 1.2217 2.0138 10.74 Decembe r 107.6883 4.4648 8.8497 39.24

The means and standard deviations of the soil erodibility, topography and crop management and vegetation factor can be found in table 3. The high value of the standard deviation is due to great variance in slope steepness. Mountainous areas have therfore higher topographic factor values than flatter areas. The maximum value of the topography factor is 71.016.

Facto Mean Standar

11

Table 2: The maximum, mean and standard deviation value, and the percentage of cumulative mean of the R-factor for each month and annually.

(12)

r d deviatio n K 0.255 5 0.0654 LS 2.536 6 2.9772 C 0.341 1 0.1625

(13)

Modelling Results

The GIS-model result can be found in figure 3. The original map can be found in appendix 3. The values in tons ha-1_year-1_{were categorized to indicate the} intensity of potential soils erosion. In table 4, the corresponding values can be found. The urban areas are included in the map because these areas are assigned with a value zero for erosion. These areas would otherwise cause a biased view on the complete image. The model result shows 10 erosion hotspots. The circles indicate gullies and the squares indicate potential erosion hotspots. Furthermore, it can be seen that the southern part of the island has more green colored areas than the northern part, which indicate lower potential soil erosion. Gullies (barrancos) have steep slopes because all water in the surrounding area flows in them and erode them further. Because of this process, they are visible in the potential soil erosion maps.

The areas within the squares are 5 large erosion hotspots, and a smaller one. Within square 2, a hilly and mountainous area can be found with sclerophyllous vegetation. Square 5 surrounds the Teide volcano. On top, bare rocks can be found while on lower altitudes coniferous forest and sclerophyllous vegetation can be found. Square 10 is an area with very steep slopes with coniferous forest

13

Figure 3: The predicted annual potential soil erosion map with marked hotspots

Figure X: The annual calculated potential soil erosion with marked potential erosion hotspots.

(14)

and sclerophyllous forest. Square 8 is a mountainous area with sclerophyllous vegetation and moors and heathland.

Category Range of values [t ha-1 year-1] Neglectable 0.00 – 0.05 Very Low 0.05 – 0.25 Low 0.25 – 1 Moderate 1 – 5 High 5 – 10 Severe 10 – 25 Very Severe 25 – 50 Highly Severe 50 – 150 Extremely Severe 150-300

In figure 4 the potential soil erosion maps of each month are shown. It is clearly visible that from April until September, hardly any potential soil erosion occurs due to lack of rainfall. In January, an erosion hotspot northeast of the Teide is visible that is not clearly present in other months. In December the potential erosion is the highest on the Teide, the area northwest of the Teide and in the northeastern corner of Tenerife.

At last, the maxima, means standard deviations and percentages of the cumulative mean for each month are shown in table 7. The similarity between the predicted soil erosion and the calculated rainfall factor is high. During the dry months (April – September) hardly any potential erosion is predicted. The difference with the rainfall factor can be found in the contribution of January and December. December contributes for 53.30% to the total predicted potential erosion, while January is here only 12.71%. The standard deviation is, except for May, June, July and September, also higher than the difference between the mean and zero. The statement that the predicted soil erosion is for the largest part of the island between zero and mean, but with big outliers, applies here too.

Month Maximum [MJ mm ha-1_h-1_year -1_] Mean [MJ mm ha-1_h-1_year -1_] Standar d deviati on % of cumulat ive mean January 30.6987 1.0644 2.4049 12.71 February 17.3811 0.6822 1.3598 8.14 March 13.1884 0.6275 0.9805 7.49 April 1.0968 0.0458 0.0812 0.55 May 0.3109 0.0036 0.0110 0.04 June 0.1311 0.0008 0.003 0.01 July 0.0022 0.0001 0.0003 0.00 August 0.0008 0.0000 0.0000 0.00 Septemb er 0.0531 0.0016 0.0038 0.02

Table 4: The categories of the legend of the erosion prediction map and the corresponding values

(15)

October 7.8784 0.2649 0.4151 3.16 Novembe r 44.1596 1.2217 2.0137 14.58 Decembe r 309.4798 4.4648 8.8497 53.30 Annual 333.7158 6.6511 10.096 -15

Tabel 7: the maxima, means, standard deviations and percentages of cumulative mean for the predicted future situation

(16)

Figure 4: Potential soil erosion on Tenerife per month, starting with January in the left top corner and ending with December in the right down corner.

(17)

Future Prospect

The results of the predicted future scenario simulation can be found in table 6 and 7. The predictions follow scenario D from the Clima Impacto, they calculated the rainfall amounts at a 2ºC temperature rise en provided maps. The simulation results in a lower mean and standard deviation for the predicted potential soil erosion. The maximum value for scenario D is higher than for the current situation. Month Maximum [MJ mm ha-1_h-1_year -1_] Mean [MJ mm ha-1_h-1_year -1_] Standard Deviatio n Current situation 333.7158 6.6511 10.096 Scenario D 342.6680 6.4700 9.4319

When comparing the results for each month from the predicted situation in table 7 with the current situation in table 5 it can be noted that the maximum values are, except for June, July and September, higher than for the current situation. In terms of means and standard deviation, only February shows a slight rise. In terms of percentages of the cumulative mean, a 3.03% increase can be seen in February while November decreases 4.38%.

Month Maximum [MJ mm ha-1_h-1_year -1_] Mean [MJ mm ha-1_h-1_year -1_] Standard Deviatio n % Of Cumulati ve Mean January 37.9279 0.8333 1.7587 13.00 February 31.5505 0.7158 1.5629 11.17 March 14.2084 0.4929 0.7405 7.69 April 1.3227 0.0252 0.0588 0.39 May 0.3399 0.0029 0.0086 0.05 June 0.1233 0.0008 0.0028 0.01 July 0.0020 0.0000 0.0001 0.00 August 0.0123 0.0000 0.0001 0.00 Septemb er 0.0351 0.0005 0.0012 0.00 October 8.2792 0.2097 0.3816 3.27 Novembe r 44.5231 0.6553 1.4413 10.20 Decembe r 326.0133 3.4734 7.1969 54.19

In figure 5, the maps for the present and future soil erosion prediction are

projected. There is hardly any visible difference. Only within the erosion hotspots, some darker spots are visible for the current situation.

17

Table 6: the maxima, means and standard deviations of the annual potential soil erosion for the current and predicted situation

Tabel 7: the maxima, means, standard deviations and percentages of cumulative mean for the predicted future situation

(18)

(19)

19

(20)

Sensitivity Analysis

A sensitivity analysis was conducted because of the uncertainty of the input values of the crop management and vegetation and the soil erodibility factor. Due to lack of local data, values from literature from other study areas had to be used. The values for both de crop management and vegetation and soil erodibility factor can be found in respectively table 8 and 9.

Land use types: Values:

 Agricultural areas o Arable land  Banana 0.21404_,0.0897_,0.128&9  Tomato 0.00 – 0.085  Vine 0.3896_{, 0.294141}7  Fruits 0.12_{, 0.1784}7 o Vine yards 0.503_{, 0.2941}7

o Fruit trees and berry plantation 0.12_{, 0.1784}7

o Pastures 0.103_{, 0.5431}7

 Forest and seminatural areas

o Broad-leaved forest 0.0031_{, 0.1302}7

 Little or no vegetation

o Sparsely vegetated areas 0.181_,0.64497

Soil type K-value Andosol 0.03-0.101 Cambisol 0.16-0.384_, 0.255 Fluvisol 0.19-0.324_, 0.33&6 Luvisol 0.40-0.444 Regosol 0.13-0.174 Solonetz 0.30-0.402

The range of possible model outcomes was analyzed to determine the impact of the uncertainty. For all land use units and soil types, the maximum calculated potential soil erosion within that unit was taken for the sensitivity analysis. For example: the cell with the highest value in figure 3 within the banana fields was found and the factor values in table 10 were the values for that specific cell. Through this method, the maximum change could be calculated. The

1_{Ranzi, R., Le, T. H., & Rulli,} M. C. (2012).

2 _{Fentie, B. et al., 2006.} 3_{Jordan, G. et al., 2005.} 4 _{Le Roux, J.J., 2005.}

5 _{Alliaume et al., 2014.} 6 _{Lieskovský & Kenderessy, 2014.} 7_{Kouli, M., Soupios, P., &} Vallianatos, F., 2009.

8_{Cox, C., & Madramootoo, C.,} 1998.

9_{Sarangi, A., Cox, C. A., &} Madramootoo, C. A., 2007.

Table 8: The land cover units and corresponding erosion values for the C-factor

1_{Rodríguez, A. et al, 2006.} 2 _{Kobza, J., & Gašová, K. (2014).}

3 _{Data from WRB (IUSS Working Group,}

2014), value calculated in excel

4_{Vopravil, J., Janecek, M., & Tippl, M., 2007} 5_{Vanelslande, A., Lal, R., & Gabriels, D.} (1987).

6 _{Vladimír, Š., Marek, K., & L'ubica, P.} (2014).

(21)

accompanying values of the other parameters can be found in tables 10 (for the C-factor) and 11 (for the K-factor). In these tables the value range is based on literature, the factors values are based on the maximum predicted value for that land use type and range result are based on calculations with the minimum and maximum value for that land use or soil type. Max/min is the maximum calculated value divided by the minimum calculated value. Class difference is the difference between legend classes of the maximum and minimum value.

Land use Value

range R-facto r K-fact or LS-facto r Range Result [t ha-1 year-1] Max/mi n [t ha-1 year-1] Class differen ce Banana 0.09 – 0.21 22.29 0.30 4.13 2.46 – 5.80 2.35 1 Tomato 0.00 – 0.08 3.76 0.31 5.28 0 – 0.49 - 2 Pastures 0.10 – 0.54 36.39 0.30 2.81 3.07 –16.57 5.40 2 Broad-leaved forest 0.003 – 0.13 2.99 0.31 12.38 0.03 –1.49 49.67 3 Sparsely vegetated areas 0.18 – 0.64 16.16 0.19 4.69 2.59 –9.29 3.59 1

The range of values from both the uncertain C-factor units and K-factor units can result in a relative big range of values for potential soil erosion. In the most extreme case of broad-leaved forest the outcome was almost 50 times higher when the largest value was used compared to the smallest one. In this case the potential soil erosion varies from neglectable to moderate following the classification table 4. Besides, tomato, pastures and cambisols have an uncertainty that differs two classes (neglectable to low). All other uncertainty values result in a difference of 1 or 0 classes.

Soil

type Value range R-facto

r LS-facto r C-facto r Range

result Max/min Classdifferen

ce Andosol 0.03 – 0.10 25.56 11.24 0.41 3.53 –11.78 3.34 1 Cambis ol 0.16 – 0.38 8.13 8.35 0.41 4.45 –10.58 2.38 2 Fluvisol 0.19 – 0.32 0.65 7.21 0.41 0.37 – 0.61 1.65 0 Luvisol 0.40 – 0.44 4.97 5.42 0.41 4.41 – 4.86 1.10 0 Regosol 0.13 – 0.17 1.77 3.83 0.79 0.69 – 0.91 1.32 0 21

Table 10: The sensitivity analysis for the land use types with an uncertainty range.

(22)

Discussion

In the discussion section, the methodology will be discussed first. After that, the results will be interpreted and the recommendations for further development will be discussed.

Methodological discussion

As stated in the method chapter, the RUSLE was originally developed for prediction soil erosion in the USA. It was also stated that the RUSLE is used for other areas around the world. Anyhow, the results of the RUSLE have to be interpreted with great care. There are multiple reasons that can lead to incorrect results.

The first reason is the theory behind the RUSLE in general. Kinnel (2008) presents various flaws of the RUSLE: dividing the slope length by 22.1 has no physical meaning, the RUSLE does not take upslope erosion into account, runoff is not generated uniformly over the hill but within the RUSLE it is assumed to be. Additionally, some remarks regarding this specific simulation can be given. Firstly, as stated earlier in this report, the values of the soil erodiblity factor and the crop management and vegetation factor had to be acquired via literature instead of field measurements. These values do apply for the same soil types or crops, but these values vary from location to location due to small differences in for instance soil composition. Besides, one value for the vegetation cover was used for the whole year, while canopy cover can vary greatly season to season. Besides, Wischmeier & Smith (1978) developed the EI30 parameter for the rainfall factor, which is the maximum raindrop intensity during a storm for 30 minutes, while in this simulation an approximation based on monthly and annual rainfall amount is used. The estimation of the rainfall amount in areas between the weather station is based on interpolation. Additionally, the negative values were set to zero. These values can therefore deviate from real values.

When the support practice factor is not taken into account, an overestimation of potential soil erosion can occur on arable lands where measures are taken to

(23)

prevent soil erosion. Unfortunately, because of absence of data on a large scale, the support practice factor is set to 1.

Results interpretation

The results of both simulations will be discussed here. First, the current situation will be analyzed. In figure 6 it can be seen that the hotspots in the northeast, northwest and the small patch just south of the Teide correspond to inferior, intermediate and superior basalt. These belong to series 1, the geological oldest part of the island. These areas belong, together with the Teide, to the steepest part of the island. Therefore, in these areas the erosion hotspots are located in Tenerife. As expected the potential erosion is lower on the south side of the island compared to the north side due to a lower amount of rainfall.

The sensitivity analysis produced satisfying results regarding the soil erodibility factor and more disappointing results for the crop management and vegetation factor. A maximum legend class difference of 1 is considered acceptable. Tomatoes, pastures, broad-leaved forests and cambisols have therefore a too large instability. The results for these areas are therefore unreliable.

As far as the future prospect simulation is concerned, the average amount of rainfall is lower than for the current situation. The maximum values are higher, which implicates higher local amounts of rainfall. This could imply higher rainfall intensity. Unfortunately, no predictions regarding the intensity could be derived from the model. The predictions are lower for the future scenario simulation while the actual erosion could enhance due to intensity increase. According to this prediction, global warming has a decreasing effect for potential soil erosion on Tenerife.

Further development

For further development the reliability of the results could be improved by improving the certainty of the values. For the rainfall factor, rainfall intensity rather than amounts of rainfall should be used. The dry side of the island could be more vulnerable to erosion (due to more sensitive soil types) than the prediction shows at the moment. The soil erodibility factor could be improved by taking soil sample from Tenerife and then calculate the values instead of use literature values. Moreover, the presence of surface crusts and rock fragments could be implemented in the model. The crop management and vegetation factor could be improved by take into account the canopy cover by for instance the use LIDAR data to calculate the normalized difference vegetation index (NDVI). Also, the presence of greenhouses on arable lands leads to an absence of potential soil erosion.

23

Figure 6: on the left the geologic map of Tenerife (Morales Matos & Pérez González, 2000) and on the right the predicted potential erosion. In appendix 5, the geological map can be found in full size.

(24)

Conclusions

Potential erosion hotspots are identified on Tenerife in areas with steep slopes such as the slopes of the Teide and the northeastern and northwestern parts of the island. The latter two belong to the geologically oldest part of Tenerife. Areas with long and steep slopes can be seen as erosion hotspots. The barancas have the longest and steepest slopes on the island and some of them can be clearly seen on the potential erosion map. Furthermore, the wetter northern part of the island is more prone to soil erosion than the drier south side, according to this simulation. The simulation results are partly influenced by the uncertainty ranges of the input variables of the soil erodiblity factor and crop management and vegetation factor. For this simulation, the soil erodiblity is less sensitive than the vegetation factor in terms of differences in potential soil erosion. At last, it can be said that the simulation results should be interpreted with great care, but that the RUSLE is a valid tool for determining erosion hotpots.

(25)

References

Agricultural Sector. (n.d.). Retrieved from http://www.todotenerife.es/en/article/show/3439/agricultura

Alejandro, M., & Omasa, K. (2007). Estimation of vegetation parameter for modeling soil erosion using linear Spectral Mixture Analysis of Landsat ETM data. ISPRS Journal of Photogrammetry and Remote Sensing, 62(4), 309-324. Alliaume, F., Rossing, W. A. H., Tittonell, P., Jorge, G., & Dogliotti, S. (2014). Reduced tillage and cover crops improve

water capture and reduce erosion of fine textured soils in raised bed tomato systems. Agriculture, Ecosystems & Environment, 183, 127-137.

Angima, S. D., Stott, D. E., O’neill, M. K., Ong, C. K., & Weesies, G. A. (2003). Soil erosion prediction using RUSLE for central Kenyan highland conditions. Agriculture, ecosystems & environment, 97(1), 295-308.

Argaman, E., Singer, A., & Tsoar, H. (2006). Erodibility of some crust forming soils/sediments from the Southern Aral Sea Basin as determined in a wind tunnel. Earth Surface Processes and Landforms, 31(1), 47-63.

Bagherzadeh, A. (2014). Estimation of soil losses by USLE model using GIS at Mashhad plain, Northeast of Iran. Arabian Journal of Geosciences, 7(1), 211-220.

Booth, B. (1973). The Granadilla pumice deposit of southern Tenerife, Canary Islands. Proceedings of the Geologists' Association, 84(3), 353-IN5.

Borley, G. D. (1974). Aspects of the volcanic history and petrology of the island of Tenerife, Canary Islands. Proceedings of the Geologists' Association, 85(2), 259-IN11.

Clima Impacto. (w.d.). Mapas mensuales de precipitación del Catálogo 4, seen on 15-6-2015, from

http://climaimpacto.eu/efectos/catalogos-climaticos/4-tenerife/variables-precipitacion/

Cox, C., & Madramootoo, C. (1998). Application of geographic information systems in watershed management planning in St. Lucia.Computers and Electronics in Agriculture,20(3), 229-250.

De Nascimento, L., Willis, K. J., Fernández‐Palacios, J. M., Criado, C., & Whittaker, R. J. (2009). The long‐term ecology of the lost forests of La Laguna, Tenerife (Canary Islands). Journal of Biogeography, 36(3), 499-514.

Diaz-Diaz, R., Loague, K., & Notario, J. S. (1999). An assessment of agrochemical leaching potentials for Tenerife. Journal of contaminant hydrology, 36(1), 1-30.

Fentie, B., Duncan, I., Sherman, B. S., Read, A., Chen, Y., Brodie, J., & Cogle, A. L. (2006). Sediment and nutrient modelling in the Burdekin NRM region.The use of SedNet and ANNEX models to guide GBR catchment sediment and nutrient target setting. Ed AL Cogle, C. Carroll and BS Sherman. Department of Natural Resources Mines and Water, Brisbane. QNRM06138.

Ferro V, Giordano G, Lovino M (1991) Isoerosivity and erosion risk map for Sicily. Hydrol Sci J 36(6):549–564

Fournier, F. (1960). Climat et érosion: la relation entre l'érosion du sol par l'eau et les précipitations atmosphériques. Presses universitaires de France.

Fu, B. J., Zhao, W. W., Chen, L. D., Zhang, Q. J., Lu, Y. H., Gulinck, H., & Poesen, J. (2005). Assessment of soil erosion at large watershed scale using RUSLE and GIS: a case study in the Loess Plateau of China. Land degradation & development, 16(1), 73-85.

Guerra, J. A., Arbelo, C. D., Armas, C. M., RODRÍGUEZ, A. R., & Mora, J. L. (2003). Erosión diferencial de Andosoles y Aridisoles en dos zonas climáticas de la isla de Tenerife. Edafología, 10(2), 229-237.

Herrera, R. G., Puyol, D. G., MartÍn, E. H., Presa, L. G., & Rodríguez, P. R. (2001). Influence of the North Atlantic oscillation on the Canary Islands precipitation. Journal of climate, 14(19), 3889-3903.

IUSS Working Group, W. R. B. (2014). World reference base for soil resources. World Soil Resources Report, 103. Jordan, G., Van Rompaey, A., Szilassi, P., Csillag, G., Mannaerts, C., & Woldai, T. (2005). Historical land use changes and

their impact on sediment fluxes in the Balaton basin (Hungary).Agriculture, ecosystems & environment,108(2), 119-133.

(26)

Kamminga, A.T., (2008). MSc Thesis 'Vulnerability assessment of potential soil erosion; A case study on Mauritius' Faculty of Science (FNWI), Master School of Life and Earth Sciences, University of Amsterdam, Amsterdam: 62p

Karydas, C. G., Sekuloska, T., & Silleos, G. N. (2009). Quantification and site-specification of the support practice factor when mapping soil erosion risk associated with olive plantations in the Mediterranean island of Crete.Environmental Monitoring and Assessment, 149(1-4), 19-28

Kinnell, P. I. A. (2008). Sediment delivery from hillslopes and the Universal Soil Loss Equation: some perceptions and misconceptions. Hydrological processes, 22(16), 3168-3175.

Kobza, J., & Gašová, K. (2014). Soil monitoring system as a basic tool for protection of soils and sustainable land use in Slovakia. Journal of Agricultural Science and Technology A, 4, 504-513.

Kouli, M., Soupios, P., & Vallianatos, F. (2009). Soil erosion prediction using the revised universal soil loss equation (RUSLE) in a GIS framework, Chania, Northwestern Crete, Greece.Environmental Geology,57(3), 483-497.

Kouli, M., Soupios, P., & Vallianatos, F. (2009). Soil erosion prediction using the revised universal soil loss equation (RUSLE) in a GIS framework, Chania, Northwestern Crete, Greece. Environmental Geology, 57(3), 483-497. Lagunilla, P. C., & Pacheco, A. H. (1987). Las erupciones históricas de Tenerife (Canarias) en sus aspectos vulcanológicos,

petrológico y geoquímico. Revista de materiales y procesos geológicos, (5), 143-182.

Le Roux, J.J., 2005. Soil erosion prediction under changing landuse on Mauritius. Dissertation submitted at the university of Pretoria, Pretoria.

Lieskovský, J., & Kenderessy, P. (2014). Modelling the effect of vegetation cover and different tillage practices on soil erosion in vineyards: a case study in Vráble (Slovakia) using WATEM/SEDEM. Land Degradation & Development, 25(3), 288-296.

Lim, K. J., Sagong, M., Engel, B. A., Tang, Z., Choi, J., & Kim, K. S. (2005). GIS-based sediment assessment tool. Catena, 64(1), 61-80.

Lu, D., Li, G., Valladares, G. S., & Batistella, M. (2004). Mapping soil erosion risk in Rondonia, Brazilian Amazonia: using RUSLE, remote sensing and GIS. Land Degradation & Development, 15(5), 499-512.

Morales Matos, G., & Pérez González, R. (2000). Gran atlas temático de Canarias. Editorial Interinsular Canaria.

Norder, S.J., (2010) BSc Thesis ‘Identifying socio-economic and bio-physical criteria for the development of sustainable land use strategies: a case study of changing land use under sugarcane on Mauritius Island' Faculty of Science (FNWI), University of Amsterdam, Amsterdam: 106p

Ranzi, R., Le, T. H., & Rulli, M. C. (2012). A RUSLE approach to model suspended sediment load in the Lo river (Vietnam): Effects of reservoirs and land use changes.Journal of Hydrology,422, 17-29.

Renard, K. G., Foster, G. R., Weesies, G. A., & Porter, J. P. (1991). RUSLE: Revised universal soil loss equation. Journal of soil and Water Conservation, 46(1), 30-33.

Rodríguez Rodríguez, A., Jiménez, C. C., & Tejedor, M. L. (1998). Soil degradation and desertification in the Canary Islands. The soil as a strategic resource: degradation processes and conservation measures”(A. Rodríguez Rodríguez et al. Eds.) Geoforma Ediciones, Logroño, 13-22.

Rodríguez, A. R., Arbelo, C. D., Guerra, J. A., Mora, J. L., Notario, J. S., & Armas, C. M. (2006). Organic carbon stocks and soil erodibility in Canary Islands Andosols. Catena, 66(3), 228-235.

Sarangi, A., Cox, C. A., & Madramootoo, C. A. (2007). Evaluation of the AnnAGNPS model for prediction of runoff and sediment yields in St Lucia watersheds.Biosystems engineering,97(2), 241-256.

Soil map of Tenerife (n.d.) Retrieved from http://geodata.science.uva.nl:8080/

geoportal/catalog/search/resource/details.page?uuid=%7B4A100238-ADE6-4A18-88B8-A43FDEAB6F3F%7D Shi, Z. H., Cai, C. F., Ding, S. W., Wang, T. W., & Chow, T. L. (2004). Soil conservation planning at the small watershed

level using RUSLE with GIS: a case study in the Three Gorge Area of China.Catena,55(1), 33-48.

Stocker, T. F., Qin, D., Plattner, G. K., Tignor, M., Allen, S. K., Boschung, J., ... & Midgley, B. M. (2013). IPCC, 2013: climate change 2013: the physical science basis. Contribution of working group I to the fifth assessment report of the intergovernmental panel on climate change.

(27)

Trimble, S. W. (1975). A volumetric estimate of man-induced soil erosion on the southern Piedmont Plateau. Present and prospective technology for predicting sediment yields and sources, 142-154.

Van der Knijff, J. M., Jones, R. J. A., & Montanarella, L. (2000). Soil erosion risk assessment in Europe (pp. 1-38). European Soil Bureau, European Commission. ISO 690

Vanelslande, A., Lal, R., & Gabriels, D. (1987). The erodibility of some Nigerian soils: A comparison of rainfall simulator results with estimates obtained from the Wischmeier nomogram. Hydrological processes, 1(3), 255-265.

Vladimír, Š., Marek, K., & L'ubica, P. (2014). Effect of Carbonates and Bivalent Cations and Their Relationships with Soil Organic Matter from the View Point of Aggregate Formation. Agriculture (Polnohospodárstvo), 60(3), 77-86. Vopravil, J., JANEČEK, M., & TIPPL, M. (2007). Revised soil erodibility K-factor for soils in the Czech Republic. Soil and

Water Conservation, 2, 1-9.

Wischmeier, W. H., & Smith, D. D. (1978). Predicting rainfall erosion losses-A guide to conservation planning. Predicting rainfall erosion losses-A guide to conservation planning

Zhang, S., Wang, Z., Zhang, B., Ren, C., Song, K., Liu, D., & Hu, L. (2010). Risk assessment and identification of critical source areas for diffuse phosphorus loss in Shuangyang River Watershed, Northeast China. Fresenius Environmental Bulletin, 19(12), 2832-2839.

(28)

Appendices

Appendix 1: The GIS RUSLE model

(29)

Appendix 2: Soil-Erodibility Values and Crop Management

and Vegetation Cover Values

Main soil

type Complex Value

Andosol 0.07

Andosol, cambisol, leptosol & regosol & rock 0.1925

Andosol, umbrisol, cambisol, andisol 0.16

Andosol, umbrisol, ferralsol, cambisol, andisol & umbrisol

0.21

Andosol & ferralsol 0.145

Andosol & leptosol 0.185

Anthrosol 0.31

Anthrosol & leptosol 0.305

Cambisol 0.25

Cambisol, luvisol & leptosol 0.323

Cambisol, umbrisol 0.275

Cambisol & leptosol 0.275

Cambisol & luvisol 0.335

Cambisol & vertisol 0.235

Fluvisol 0.3

Leptosol 0.3

Leptosol & regosol 0.225

Rocks & leptosol 0.3

Luvisol 0.42

Luvisol & leptosol 0.36

Regosol 0.15

Solonetz 0.35

Solonetz, solonchak, cambisol , calcisol &

leptosol 0.26

Solonetz, solonchaks, cambisol, calcisol &

anthrosol 0.2625

Solonetz, solonchak, cambisol, calcisol 0.2467

Urban 0

Vertisol 0.22

Vertisol & antrosol 0.265

Vertisol & leptosol 0.26

(30)

Land use types: Values:  Urban areas:

o Continuous urban fabric 0.0

o Discontinuous urban fabric 0.0

o Industrial or commercial units 0.0

o Port areas 0.0

o Airports 0.0

 Non-vegetated semi-urban areas

o Mineral extraction sites 0.52

o Dump sites 0.512

o Construction sites 0.0

o Sport and leisure facilities 0.0

 Agricultural areas o Arable land  Banana 0.21404_,0.0897_,0.1210&11  Tomato 0.00 – 0.085  Potato 0.429  Vine 0.3896_{, 0.294141}8  Ochard 0.203  Fruits 0.12_{, 0.1784}8 o Vine yards 0.503_{, 0.2941}8

o Fruit trees and berry plantation 0.12_{, 0.1784}8

o Pastures 0.103_{, 0.5431}8

 Forest and seminatural areas

o Broad-leaved forest 0.0031_{, 0.1302}8

o Coniferous forest 0.33808

o Moors and heathland 0.50018

o Sclerophyllous vegetation 0.41288

o Traditional woodland-scrub 0.37558

 Little or no vegetation

o Beaches dunes sands 0.57548

o Bare rocks 0.78508

o Sparsely vegetated areas 0.181_,0.64498

1_{Ranzi, R., Le, T. H., & Rulli,} M. C. (2012).

2 _{Fentie, B. et al., 2006.} 3_{Jordan, G. et al., 2005.} 4 _{Le Roux, J.J., 2005.}

5 _{Alliaume et al., 2014.} 6 _{Lieskovský & Kenderessy, 2014.} 7_{Angima, S. et al., 2003.} 8_{Kouli, M., Soupios, P., &} Vallianatos, F., 2009. 9_{Shi, Z. et al., 2004.}

10 _{Cox, C., & Madramootoo, C.,} 1998.

11 _{Sarangi, A., Cox, C. A., &} Madramootoo, C. A., 2007.

12_{Karydas, C. G., Sekuloska, T., &} Silleos, G. N. (2009).

(31)

Appendix 3: Potential soil erosion maps

(32)

(33)

(34)

(35)

(36)

(37)

(38)

(39)

(40)

(41)

(42)

(43)

(44)

Appendix 5: The Geological Map of Tenerife

(45)

Potential Soil Erosion Simulation On Tenerife With Use Of The Revised Universal Soil Loss Equation

June

2015