Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Guidelines for Monitoring and assessment of wind erosion at site level Prepared for: Arab Center for the Study of Arid Zones and Dry Lands (ACSAD) Damascus, Syria By: Dhruba Pikha Shrestha May 2008 International Institute for Geo-information Science and Earth Observation (ITC) P. O. Box 6, 7500 AA Enschede, the Netherlands 2 3 Executive summary Land degradation is a worldwide issue. The estimated surface area affected by human-induced soil degradation is 24 per cent of the inhabited land areas (Oldeman, 1994; Oldeman et al., 1991). Of various degradation processes soil erosion by action of wind plays an important role in semi arid and arid environment. The causative factors are the dryness of the environment, soil properties and vegetation cover. Wind erodes finer sediments by a process called deflation. The sediments are transported by saltation and/or suspension. The wind transported sediments can also wear away rocks. This process is called abrasion. When there is selective removal of finer sediments by wind, there remains surface coarser sediments which are called desert pavements. The principal depositional landform of wind action is dune formation. Depending on sediments supply, wind characteristics and vegetation cover several types of dune formation are possible. For monitoring and assessing wind erosion the GLASOD approach would be very much useful. For assessing erosion features air photo interpretation technique helps in delineating accurately the type, extent and severity of wind erosion patterns. In case of unavailability of recent air photos, a rapidly deployable and effective low cost method can be applied to detect and assess wind erosion damage in its early stages. Photographs can be taken from a small aircraft by carefully planning flight lines to cover the study area. If photos are not digital, they can be scanned and converted into raster data format to be used in a GIS system. For geo-referencing the photos ground control points can be taken using a GPS receiver. Satellite data will be useful to map large areas and also to derive indices useful to map erosion hazard. In addition, the combination of remote sensing data and digital elevation model helps generate stereoscopic view of the area which can be used for delineating erosion features at large areas efficiently. For applying wind erosion model, one has to clearly define the objective of assessment and also to check if required data for running the model is available. A suitable model can then be selected. In addition, it is very important that field data is available for calibration and validation of the model results. 4 The case study included towards the end of the guidelines is on wind erosion feature mapping and monitoring, based on the work of Hennemann and Nagelhout (2002). The idea is that in case of unavailability of conventional aerial photos or if they are too old, small format aerial photography (SFAP) can be taken aboard a small aircraft and used to map and monitor erosion features. The case study consists of 2 parts: the analysis of the performance of SFAP and the analysis and assessment of wind erosion. The use of SFAP proved in spite of some minor operational drawbacks to be quite promising in terms of geometric accuracy, visual quality, cost effectiveness and above all timeliness. Generation and application of SFAP allowed for a detailed assessment of the current erosion status followed by a spatio-temporal GIS analysis of wind erosion patterns and trends in relation to key environmental factors. 5 I. Introduction Desertification is the degradation of land in arid, semi arid and dry sub-humid areas resulting primarily from human activities and influenced by climatic variations. It occurs predominantly, but not exclusively, in semi-arid areas (Eswaran et al., 2001). A major impact of desertification is loss of biodiversity and loss of productive capacity of land. It is also associated with change of vegetation e.g. from perennial grasses to one dominated by shrubs. Primary reasons for desertification are overgrazing, over cultivation, deforestation, overdrafting of groundwater and global climate change. While drought is a contributing factor, the main causes are related to man’s overexploitation of the environment. In drier environment land degradation by wind action is very crucial. Winds may erode, transport and deposit materials, and are effective agents in regions with sparse vegetation cover and a large supply of unconsolidated sediments. Wind erosion process, also known as Aeolian process involves the removal of loose and fine-grained particles including organic matter from the surface of Earth, their transportation by various processes, and finally the deposition of the particles. The effect of wind erosion is fertility depletion in agricultural fields leading to reduction in crop harvest and desertification in the long run. The offsite effects of wind erosion can be accumulation of sand and dust on the fields, drainage ditches, farm machinery, surface water, infrastructures such as roads, railways, buildings etc. In a global extent, wind erosion accounts for about 28 % of area affected by land degradation (Oldeman, 1994). Wind action in erosion, transport and subsequent deposition of fine particles has been recognized as an important environmental problem (Goossens and Riksen, 2004). Coarse and finer particles enter the atmosphere through various mechanisms, affecting a large number of physical and chemical processes and consequently the natural environment. In drier regions of the world this has been major environmental issue. Wind action not only affects erosion and deposition of soil particles but the atmospheric dust also causes environmental pollution. The concentration of dust in the atmosphere also influences the climate. The short term effects of high dust concentration in the atmosphere is reduction of visibility, this is especially the case during dust storms. If 6 pesticides are used in agricultural fields this can be harmful to the surrounding areas. The long term effects are due to the transportation of finer dust particles which may carry organic matter, heavy metals, pesticides and fertilizers long distances from their source. This will have negative effects on human health. The effects of fine airborne particles on environmental pollution have been a subject of study, in the field of both rain and water pollution. It is reported that fine atmospheric dust affects various aspects of human health (Griffin et al., 2001; Smith and Lee, 2003). In addition, the role of fine dust particles in the atmosphere in climate change can not be underestimated. This will certainly be the topic of many researchers. II. General setting of wind action in drylands II. 1 Processes and types Like water, wind also erodes sediment more readily. Wind is also an agent of sediment transport which moves soil particles by rolling them along the surface, by the processes known as saltation and/or suspension. The finest particles are suspended in the air and carried long distances. Saltation is a process by which sheets of sand (size between 50 – 500µ) is raised by violent wind and is transported few meters over smooth surfaces, leaving sheets of sand on the ground or small mounds of sand trapped by plant (Figure 1). These sand sheets can cause serious crop damage. Wind exercises a pressure on solid particles in repose. This pressure is exerted above the centre of gravity on the surface exposed to wind and is opposed by a friction centred on the base of the particles. Like water, wind erodes sediment more readily than solid rock. Abrasion is a process by which wind transported sediment can wear away at rock. Wind abrasion is a sort of natural sandblasting, very similar to milling by sand-laden waves. Abrasion during dust storms is capable of eroding rocks. Since wind-borne sand is seldom lifted far off the ground, wind abrasion is generally a near surface process. Deflation is the removal of unconsolidated sediment by wind action. In this process light and 7 finer particles of soil (less than 100 µ diameter, e.g. clay, silt and fine sand including organic matter) are carried away in suspension (Figure 2). This dust is sucked up by wind and carried several kilometres and then dispersed as a dry mist, or it may travel several hundred kilometres as a dust cloud. Deflation is most active where winds are unobstructed and the sediments exposed unprotected by vegetative cover. Figure 1: Sand movement by a process called saltation 8 Figure 2: Fine soil particles in suspension Wind related products are called eolian. Eolian sediments are well sorted in terms of size since it is related to wind velocity. Winds selectively move finer particles; as they slow down they drop coarser particles first. When there is selective removal of finer sediments by wind and seasonal surface runoff, there remains surface sediments of residual coarser materials called desert pavement (Figure 3). This desert pavement then protects underlying finer sediments from further erosion. Desert pavement surface, once established, can be very stable. It is also known as deflation armour. 9 Figure 3: Desert pavement The principal eolian depositional landform is the dune, a low mound or ridge of sediment. Dunes begin to form when sediment-bearing winds encounter an obstacle that slows them down. With reduced velocity, the wind begins to drop the coarsest or heaviest fraction of its load. The deposition, in turn, creates a larger obstacle creating a windbreak, causing more deposition in a self –reinforcing cycle. Dunes can have height of 3 – 100 m. Dunes are not a permanent or static objects. Once formed, they tend to migrate if wind continue to blow predominantly from a single direction. There are several types of dunes, depending on sediment supply, wind characteristics and vegetation. The transverse dunes are elongated perpendicular to the prevailing wind direction. Many of these dunes have a crescent shape, with arms of horns pointing downward. They are called barchan dunes (Figure 4, Figure 5). Barchan dunes are known to form in regions with limited sediment supply and unidirectional wind regimes (Daniell and Hughes, 2007). 10 Figure 4: Barchan dune Figure 5: Barchan dunes as seen on a satellite image. The longitudinal dunes occur where sediment supply is limited and winds are relatively strong. These dunes are elongated and parallel to the direction of wind flow (Figure 6). 11 Figure 6: The parallel longitudinal dunes II. 2 Impacts of wind erosion Effect of wind erosion can be on-site as well as off-site. The on-site effects are loss of topsoil and plant nutrients, which have direct impact on crop growth. Soils become less productive because they contain less nutrients and less capacity to retain water. A field experiment conducted in the effect of wind erosion and sand accumulation in inner Mongolia shows that long term wind erosion could result in significant soil coarseness, infertility and dryness (Zhao et al., 2006). Abrasion caused by flying soil particles does considerable damage to crops and to young plants in particular. In addition to this, evaporation from plant leaves is accelerated by 12 wind, restricting wheat growth. The off-site effects are due to sand cover on fertile agricultural areas which affects crop growth and eventual decrease of harvest. In number of situations there will be soil textural changes resulting in decrease of clay particles and reduction in the ability of soil to conserve water. In a study of the effect of wind erosion on soil properties in China, similar results are reported: decrease of clay content and nutrient reduction in the soil e.g. decrease of organic matter, nitrogen and phosphorus contents (Jian et al., 1992). Also infrastructures can be covered by over-blown sand which will be very nuisance (Figure 7). In extreme cases the land becomes useless because of thick sand cover (Figure 8). Fine dust in the atmosphere will have environmental problem causing health hazard to human beings. Figure 7: Sand covering part of a main road in Priai, Cape Verde 13 Figure 8: Over blowing with dune formation, case of extreme wind action II. 3 Distribution 3.1 Global context Desertification is a worldwide issue. It affects all the continents of the world (Figure 9). However, the present situation shows that it is predominantly affecting large parts of the world between the latitudes 15 – 35o N, which are relatively hotter and drier. But wind action not only limits to this zone, large part of Australia, Mongolia, China, Europe, United States of America and many countries in south America are affected. In drier and open environment where there are no obstructions soil degradation by wind action is very crucial. After water erosion, wind action in erosion is very important degradation process. The global estimates of area affected by wind erosion are in the order of 5.49 million km2 (Table 1) which occurs mainly in drier regions. 14 Table 1: Estimates of the global extent (in million km2) of land degradation (Oldeman, 1994) Type Light Moderate Strong +extreme Total Water erosion 3.43 5.27 2.24 10.94 Wind erosion 2.69 2.54 0.26 5.49 Chemical degradation 0.93 1.03 0.43 2.39 Physical degradation 0.44 0.27 0.12 0.83 Total 7.49 9.11 3.05 19.65 Figure 9: Desertification vulnerability (Eswaran et al., 2001) 3.2 Arab context The Arab world (Figure 10, Table 2) stretches over an area of 13551029 Km², divided in the Arab Maghreb (Algeria, Libya, Morocco, Mauritania and Tunisia), the Arab Mashreq (Jordan, Lebanon, Syria Arab Republic and Iraq), the Central region (Egypt, Sudan, Djibouti and Somalia), and the Arabian Peninsula (Kuwait, Saudi Arabia, Bahrain, Qatar, United Arab Emirates, Oman and Yemen) (ACSAD, 2007). With the exception of Bahrain, Kuwait, Lebanon and Syria population density is less than 100 persons per Km2 (Table 2). 15 Apart from Yemen and some areas in Jordan and Syria, annual rainfall in most of the area is less than 400 mm. Also vegetation cover is sparse. The analysis of the satellite vegetation images (NDVI images) of the Arab World shows negative change of vegetation cover during the period 1999-2005 in all the countries of the Arabian Peninsula (Table 3) (ACSAD, 2007). Similar trend is reported in Libya, Djibouti and Lebanon, indicating the threat of land degradation problem in many Arab countries. Table 2: Arab world (ACSAD, 2007; UNEP and Leauge_of_Arab_States, 2004) Country United Arab Emirates Jordan Bahrain Tunisia Algeria Djibouti Saudi Arabia Sudan Syria Somalia Iraq Oman Palestine Qatar Kuwait Lebanon Libya Egypt Morocco Mauritania Yemen Surface area (Km2) 82880 92300 665 163610 2381740 23000 1960582 2505810 185180 637657 437072 212460 62206 11437 17820 10400 1759540 1001450 446550 1030700 527970 Population 4496000 5703000 727000 10102000 32854000 793000 24573000 36233000 19043000 8228000 28807000 3100300 NA 813000 2687000 3577000 5853000 74033000 31478000 3069000 20975000 Pop. density (persons/Km2) 54 62 1093 62 14 34 13 14 103 13 66 15 NA 71 151 344 3 74 70 3 54 Annual rain (mm) 65 150-600 80 NA 250 30 50-100 NA NA 50 – 200 300 – 1200 50 - 100 150 - 300 75 31 - 318 NA NA 40 – 75 NA NA 1000 Soil degradation by wind action is thus common in most of the countries. In Lebanon, however, water erosion seems to be the main issue. In the map of the global assessment of human induced soil degradation (GLASOD) the area covered by active dune formation is quite extensive in Algeria, Morocco, Tunisia, Libya, Egypt, part of Sudan, Saudi Arabia, Yemen, Oman and part of United Arab Emirates. Similarly, in considerable areas of these countries wind action is 16 responsible for the removal of topsoil resulting in nutrient depletion leading to desertification. It is reported that desertification in north-western Jordan is taking place through overgrazing, erosion, soil fertility depletion and decreased productivity. Erosion by wind and water is considered to be the major cause of land degradation in Jordan (Khresat et al., 1998). The soils contain little organic matter and their alkaline reactions reduce the availability of phosphorous and macronutrients and consequently lead to very low crop yields. In Syria evidence of wind erosion is widespread in the drylands as shown in a case study from the Khanasser valley (Masri et al., 2003). Similarly in the Sudan wind erosion is the most widespread soil degradation type, especially in the arid zones (Ayoub, 1998). Figure 10: The Arab extent (ACSAD, 2007) 17 Table 3: Distribution of positive and negative vegetation change during 1999-2005 in the Arab countries (ACSAD, 2007) II. 4 Causative factors The main causative factor is the dryness of the environment. In areas where annual rain is less than 600 mm there are more than 6 months of dry months without rain. Potential evapotranspiration rate is also very high in the dry regions (more than 2000 mm). In these cases soils are often bare for long period because of lack of water in agricultural areas. If there is strong wind the chance is much higher that there will be erosion. Other factor is related to soil properties. If soil particle size is dominantly between 10 and 100 microns in size, it is very vulnerable. Clayey soils are more resistant and coarse sand particles are too heavy to be removed by wind. Soil organic matter content is equally important since it helps in maintaining good soil structure. When soil is moist it is also resistant since it helps in 18 cohesion. Vegetation cover is equally important. If the area is covered by vegetation it protects the soil against wind action in the same way it protects against the impact of raindrops. III. Methodology III.1 Monitoring wind erosion For monitoring wind erosion it is necessary to assess current status of the problem and assess risk which may occur in future. Assessment of the current status reflects what has happened to date. It is assessed by direct observation and expert judgement. To predict a potential situation that may occur in future a risk assessment can be made. Risk assessment is generally made based on modelling. For mapping purposes various techniques can be applied such as the use of remote sensing techniques and the use of models. For monitoring wind erosion the approach described in the global assessment of human induced soil degradation (GLASOD) (Oldeman et al., 1991) will be very suitable. In the GLASOD methodology three wind erosion features are recognised: loss of topsoil (Et), terrain deformation (Ed) and over-blowing (Eo). Loss of topsoil by wind action is a widespread phenomenon in arid and semi-arid environment. In general coarse textured soils are more susceptible to wind erosion than fine textured soils. In (semi-)arid environment natural wind erosion is often difficult to distinguish from human-induced wind erosion, but natural wind erosion is often aggravated by human activities. Terrain deformation by wind erosion is a much less widespread type of degradation than loss of topsoil. It is defined as the uneven displacement of soil material by wind action and leading to deflation hollows and dunes. It can be considered as an extreme form of loss of topsoil, with which it usually occurs in combination. Overblowing is defined as the coverage of the land surface by wind-carried particles. It is an off-site effect of the wind erosion types mentioned above. When it is at extreme case the whole area is then be covered by sand. Over blowing may occur in the same mapping unit as those other types, or in adjacent units. It may influence structures like roads, buildings and waterways but it can also cause damage to 19 agricultural land. Stereoscopic air photo interpretation can be used to delineate boundaries of erosion features. In addition to mapping wind erosion features degree of degradation can be assessed. The degree of degradation can be categorized into 4 classes as follows: (1) Light: if the terrain has somewhat reduced agricultural suitability, but it is suitable for use in local farming systems. Restoration to full productivity is possible by modifications of the management system. Original biotic functions are still largely intact. (2) Moderate, if the terrain has greatly reduced agricultural productivity but is still suitable for use in local farming systems. Major improvements are required to restore productivity. Original biotio functions are partially destroyed. (3) Strong, if the terrain is not reclaimable at farm level. Major engineering works are required for terrain restoration. Original biotic functions are largely destroyed and (4) Extreme, if the terrain is not reclaimable and beyond restoration. Original biotic functions are fully destroyed. It is also useful to indicate whether the degradation type recognised in a mapping unit (geomorphic mapping unit?) occurred (1) infrequently (up to 5% of the unit), (2) common (610% being affected), (3) frequent (11-25% being affected), (4) very frequent (26 – 50% being affected) or (5) dominant (more than 50% affected). In addition, it would be also useful to indicate the causative factors. Causative factors could be over grazing, over exploitation of land, etc. Mapping symbol could be combination of alphanumeric letters e.g. Eo2.3g (Eo = Overblowing, 2 = moderate degree of degradation, 3 = frequent or 11-25% of mapping unit affected, and g = overgrazing as being the cause of accelerated erosion). III.2 Assessment of wind erosion 2.1 Use of Aerial Photography Assessment of wind erosion can be made based on image interpretation. Air photos can be 20 analysed stereoscopically to delineate areas of wind erosion features e.g. areas affected by loss of topsoil, active dune formation, areas affected by over blowing, etc. If aerial photos are too old the new wind erosion features may not be possible to map. In such a situation a rapidly deployable and effective low cost method to detect and assess wind erosion damage in it early stages should be an alternative method. With such a method one should be able to: - accurately establish the type, spatial extent, severity and spatio-temporal trends of the wind erosion pattern in the source areas; - analyse the nature of the underlying erosion processes and causal factors involved; - achieve the above within acceptable geometric accuracy limits and with due effectiveness and efficiency in terms of manpower, cost and time. One of the alternate and rapid methods of assessing wind erosion damage is the use of smallformat aerial photography (SFAP). SFAP has proven to be valuable instrument for various resource inventory studies such as in forestry, rangeland mapping, etc. Tueller et al. (1988) used SFAP to measure changes in rangeland vegetation in north-eastern Nevada, USA using helicopter-borne 35 mm aerial photography. Hennemann and Nagelhout (2004) applied SFAP in Naivasha area in the Rift valley, Kenya as a tool to develop wind erosion detection and assessment method. In the Naivasha area in Kenya the SFAP photographs were taken with a 35 mm Minolta camera from a small Cessna 182 aircraft in the year 2000; a total of 130 photographs were shot in 4 flight lines to cover the study area. Sufficient ground control points were taken to rectify the photos. The photos were scanned and imported as raster data in a GIS system (ILWIS 3.4). Using old aerial photos from 1991 of the area erosion severity maps of the two different periods, general trends and rates of land degradation for the period 1991-2000 could be calculated and assessed. The use of small-format aerial photography applied in the Naivasha area, Kenya shows that an up-to-date colour photo cover of the entire study area can be generated at a scale of 1:5,000 at a reasonable cost and that erosion severity map can be generated quickly. 21 2.2 Use of satellite remote sensing techniques Satellite image interpretation can be done monoscopically on colour composites at relevant scale. For making false colour composites the band combinations of Landsat TM band 4 (red), TM band 5 (green) and TM band 3 (blue) is commonly used. In case of ASTER and SPOT image the ideal combination would be band 3 (red), band 2 (green) and band 1 (blue). Once the colour composites are made interpretation can be carried out for wind erosion features following GLASOD approach. If stereo pair of satellite data is available e.g. SPOT stereo pair or ASTER stereo pair interpretation can be carried out using stereoscope or directly on compute screen. Stereo pair can be also generated if digital terrain model (DTM) or contour data is available. To generate stereoscopic image from the combination of satellite image and DTM, a GIS system e.g. ILWIS can be applied. The procedure is explained in detail as follows: For stereoscopic interpretation of satellite image a stereo pair is necessary. In case of unavailability of stereo pair (e.g. stereo SPOT image or stereo ASTER image) digital elevation model (DEM) can be also used. For this purpose a GIS software is indispensible. The ITC developed GIS software (ILWIS 3.4) can be freely downloaded from http://52north.org/index.php?option=com_content&task=view&id=131&Itemid=155. ILWIS is a remote sensing and GIS software which integrates image, vector and thematic data in one unique and powerful package on the desktop. ILWIS delivers a wide range of features including import/export, digitizing, editing, analysis and display of data, as well as production of quality maps. ILWIS software is renowned for its functionality, user-friendliness, and has established a wide user community over the years of its development. In ILWIS the stereo pair from DTM operation creates a stereo pair from a single raster image and a Digital Terrain Model of the area (Figure 11). This stereo pair can be viewed on the computer screen (split screen mode) with a stereoscope (screen-scope) or as an anaglyph using red-green or red-blue glasses (Figure 12). To create a stereo pair in ILWIS following steps need 22 to be followed: a. The satellite image has to be first imported into ILWIS format. The image will have to be georeferenced selecting an appropriate coordinate system. The parameters required for georeferencing can be derived from the topographic map of the area. The DTM has also to be imported into ILWIS format which should have the same geo-reference. DTM can be also created from digitized contour lines by interpolation. It is also possible to use an existing DTM such as the SRTM DTM or an ASTER DTM. Figure 11: Generation of stereo pair from DTM and Landsat TM data b. In ILWIS use the “Stereo Pair from DTM” option from the operations list and select the georeferenced raster image (satellite image). Also select the DTM of the same area as the image. It 23 is important that image and DTM have (approx.) the same resolution. The operation will create a stereo pair which can be displayed and viewed using a special stereoscope which can be mounted on computer screen. The stereo pair can be also displayed as an anaglyph for which red-green or red-blue glasses will be necessary to view the image. Detail information on various options for creating stereo pair from DTM in ILWIS is given in the help function of ILWIS. This is shown in the Table 4. Apart from visual interpretation for wind erosion features, satellite data can be also used in computing vegetation canopy estimation which can be used for wind erosion hazard assessment. Canopy cover is generally computed using Normalized Difference Vegetation Index (NDVI). A linear function to derive USLE C- factor in a case study in Southern France with the resulting equation of C = 0.431 – (0.805xNDVI) is described by de Jong (1994). For estimation of canopy cover exponential function seems to give better result than linear function (Van der Knijf et al., 1999). Exponential function is given by: C=e −α ( NDVI ) ( β − NDVI ) (1) where α and β are the parameter determining the shape of the curve with constants 2 and 1 used for α and β respectively. 24 Figure 12: Stereo anaglyph (a) and image interpretation based on the anaglyph (b) (Shrestha et al., 2005) 25 Table 4: Various options for creating stereo pair in ILWIS Input raster map: Select an input raster map, i.e. a single scanned aerial photograph, a satellite image, or a normal raster map. The map should use the Image domain, any other value domain, the Color domain, or the Picture domain. Open the list box and select the desired input map, or drag a raster map directly from a Catalog into this box. On the commmand line, there are no restrictions on the domain type of the input map. DTM: Select an input Digital Elevation Model (DTM) (raster value map). If the DTM does not cover the entire area of the input raster map, then the parts of the input raster map for which no DTM value can be found will appear black in the output stereo pair. A Digital Terrain Model is also known as a Digital Elevation Model (DEM). Tip: It is highly recommended to have the Interpolation check box selected in the Properties sheet of your DTM. Look angle: Specify a value between 0 and 90° for the total angle of 'projection' of the two output raster maps which together will form the output stereo pair. For more information, refer to Stereo pair from DTM : functionality. Reference height: Specify a height value (of your DTM) that should appear at 'ground' level of your monitor when inspecting the stereo pair. Larger height values in the DTM will appear 'outside of your monitor'; smaller height values in the DTM will appear 'inside your monitor'. Look modus: Select how the Look angle should be divided over the two output raster maps that will form the stereo pair. Left: The left output raster map will be 'projected' using the total Look angle; the right output raster map will be 'projected' vertically. Both: The left and the right output raster map will each be 'projected' using half the Look angle. Right: The left output raster map will be 'projected' vertically; the right output raster map will be 'projected' using the total Look angle. Resample modus: Select the method to resample the input raster map into two output raster maps which together will form the output stereo pair. Fast: For quick resampling. Resampling is input-driven (the estimated position of output pixels is less accurate) and biased nearest neighbour. Accurate: For slow but accurate resampling. Resampling is output-driven just as the Resample operation; value maps and maps with the Color domain are resampled using bilinear interpolation; other maps are resampled using nearest neighbour. Output stereo pair: Type a name for the stereo pair. The (resampled) output raster maps of the stereo pair will also obtain this name, followed by _Left or _Right. 26 2.3 Application of model Modelling wind erosion started in the early sixties for semi-quantitative estimation of soil losses. Wind erosion modelling has been mainly semi-empirical and focussed mainly on on-site effects. The Wind Erosion Equation (Woodruff and Siddoway, 1965) is a good example. It has been applied in several locations (Klik, 2008; Panebianco and Buschiazzo, 2008; Wassif et al., 2002). Klik (2008) reports that soil erosion assessment by wind erosion equation linked to a GIS enables the designation of potential risk areas and that the results seem to be reasonable. The Wind Erosion Equation (WEQ) calculates potential average annual erosion rates as follows: E = f ( I , K , C , L, V ) (2) Where, E is the potential annual soil loss (t ha-1 yr -1), I is the soil erodibility, expressed as potential annual soil loss in (t ha-1 yr -1) from a wide, unsheltered isolated field with bare, smooth, level, loose and non-crusted surface, K is the surface roughness factor which is a measure of the effect of ridges made by tillage and planting implements, or other means of creating systematically spaced ridges. Ridges absorb and deflect wind energy and trap moving soil particles. C is an index of climatic erosivity, specifically wind-speed and surface soil moisture. The factor for any given location is based on long-term climatic data and is expressed as a percentage of the C factor, L is the unsheltered, weighted travel distance (in m) along the prevailing wind direction, V is the equivalent vegetation cover expressed by relating the kind, amount, and orientation of vegetative material to its equivalent in kg ha-1 of small grain residue in reference condition (SGe). The equation includes some functional relationships which are not linear mathematical calculations. 27 Soil erodibility factor I The soil erodibility factor I is related to the percentage of non-erodible surface soil aggregates larger than 0.84 mm in diameter which can be determined by dry sieving method (Chepil, 1942). Based on the soil texture 8 Wind Erodibility Groups are distinguished. A distinction is made in calcium carbonate content because soils high in CaCO3 (> 5%) are more erodible. The Wind Erodibility Groups have been applied to the Austrian Texture Triangle (Figure 13). The soil erodibility index was then calculated as weighted average of the area of each texture class (Table 5). Soil erodibilty factor is also affected by topographic features like knolls. Knolls are topographic features characterized by short, abrupt windward slopes. Wind erosion potential is greater on knoll slopes than on level or gently rolling terrain because wind flow-lines are compressed and wind velocity increases near the crest of the knolls. Erosion that begins on knolls often affects field areas downwind. Adjustments of the soil erodibility index I are used where windwardfacing slopes are less than 160 m long and the increase in slope gradient from the adjacent landscape is 3 percent or greater. Both slope length and slope gradient change are determined along the direction of the prevailing erosive wind. Figure 13: Application of wind erodibility groups to the Austrian soil texture triangle (Klik, 2008) 28 Table 5: Soil erodibility index (t ha -1 yr -1) for soil textural classes Surface roughness factor K The roughness factor K describes the effect of soil surface roughness on soil erosion. It is distinguished between random roughness (Allmaras et al., 1996) and oriented roughness made by tillage and planting implements, or other means of creating systematically spaced ridges. Kfactor for oriented roughness can be also determined using equations by Williams (1986) where ridge height and ridge distance in prevailing wind direction are considered. The random roughness values used in the WEQ are the same values used in the Revised Universal Soil Loss Equation RUSLE (Renard et al., 1997). The effect of random roughness on K is only used with the Management Period Procedure. The random roughness factor accounts for roughness effects on soil erodibility. It considers that surface roughness of soils with high erodibility decreases faster than of less erodible soils. Climatic factor C The C factor is an index of climatic erosivity, specifically wind-speed and surface soil moisture, and is expressed as a percentage of the C factor, which has been assigned a value of 100 (Lyles, 1983). The climatic factor equation is expressed as: C = 386.u 3 /( PE ) 2 (3) 29 Where, C is the annual climatic factor, u is the average annual wind velocity, PE is the precipitation-effectiveness index of Thornthwaite, which is calculated by: PE = 3.16.Σ[ Pi /(1.8Ti + 22)]10 / 9 (4) Where, Pi is monthly precipitation in mm and Ti average monthly air temperature in °C. The prevailing wind erosion direction is the direction from which the greatest amount of erosive wind occurs during the critical wind erosion period or time period being evaluated. Beside wind speed the erosive wind energy (EWE) has the main impact on the erosion process. When hourly wind speed data are available the hourly EWE can be assessed by following equation: EWE hr = 3600. p.U 2 (U − Ut ) (5) Where, EWEhr is the hourly erosive wind energy in g s-1, p is the air density (g m-3 ), U and Ut are the average hourly wind speed and average hourly threshold wind speed (m s-1), respectively. For the threshold wind speed a value of 8 m s-1 is often indicated. Unsheltered distance L The L factor represents the unsheltered distance along the prevailing wind erosion direction for the field or area to be evaluated. It is the total length of the field reduced by the length sheltered 30 by protection measures. The determination of the unsheltered distance is done stepwise: - Determination of an isolated field - Determination of wind breaks and their properties - Determination of prevailing wind direction - Calculation of sheltered field length by wind breaks - Calculation of unsheltered distance The L factor begins at a point upwind where no saltation or surface creep occurs (stable) and ends at the downwind edge of the area being evaluated. The investigation area had to be divided into a number of fields which could be considered as isolated from each other. This means that no soil particles are crossing these field boundaries. Field boundaries consist of wind breaks, roads and field paths, creeks etc. For the determination of isolated fields topographical maps can be combined with information from satellite images and aerial photos. Satellite imageries or aerial photos can be useful for the knowledge of the field geometry, tillage direction, existence of wind breaks and roads. Determination of area sheltered by wind breaks can be estimated as a length of 15 times the height of the wind break assumed in prevailing wind direction (Tibke, 1988). However effectiveness of the wind break depends on the porosity of the wind breaks (Table 6). Table 6: Sheltered field length Porosity Sheltered field length (x height of wind breaks) in m Low 15 Medium 10 High 5 Very high 2 Vegetation cover factor V The effect of vegetative cover in the Wind Erosion Equation is expressed by relating the kind, 31 amount, and orientation of vegetative material to its equivalent in kg per hectare of small grain residue in reference condition (SGe). This condition is defined as 25 cm long stalks of small grain, parallel to the wind, lying flat in rows spaced 25 cm apart, perpendicular to the wind. Position and anchoring of residue is important. In general, the finer and more upright the residue, the more effective it is for reducing wind erosion. (Williams et al., 1990) proposed an equation for the V-factor based on the small grain equivalent SGe: V = 0.2533( SGe)1.363 (6) The SGe can be assessed by: SGe = g1.BAG + g 2. SR + g 3. FR (7) where g1, g2, g3 are crop coefficients, BAG is the above ground living biomass (kg ha-1), SR is the standing residues (kg ha-1) and FR is the flat residues (kg ha-1). The variables g1 - g3 are derived from Williams et al. (1990) and are shown in Table 7 for the main crops. Table 7: Coefficients g1, g2, g3 for calculation equivalent SGe for main crops Crop g1 g2 Summer barley 3.390 3.400 Winter wheat 3.390 3.400 Sugar beet 1.140 0.600 Soybean 1.266 0.633 Corn 0.433 0.433 Sorghum 0.657 0.657 Summer wheat 3.390 3.390 Sunflower 3.390 3.390 Potatoes 3.390 3.390 of small grain g3 1.510 1.610 0.330 0.729 0.213 0.320 1.610 1.610 0.320 The protective impact of the vegetation depends also on the angle between prevailing wind direction and the tillage direction or direction of planting. In the WEQ a correction factor considers this fact. For the Management Period Procedure the knowledge of temporal distribution of living biomass production, residue amounts and impact of tillage on soil surface 32 roughness is necessary. Depending on availability of detail data including wind velocity rates at a given storm event a dynamic process-based model such as the Texas Tech Erosion Analysis Model (TEAM) can be also applied to predict detachment, movement and deposition of soil particles associated with wind processes (Gregory et al., 2004). TEAM simulates the suspension and movement of dust above and downwind from eroding sites and predicts a horizontal movement in mass per unit width per unit of time for wind and humidity inputs. IV. Conclusion and recommendation Air photo interpretation is very useful technique for mapping wind erosion features. For assessing the effects of wind action, GLASOD method will be very much suitable. The method may need some adjustments to make it suitable for assessment at local scale. The drawback of air photos is that they will become soon old and it will not be possible to map recent erosion features. If there is no financial constrains high resolution image (e.g. IKONOS image) can be obtained for monitoring wind erosion. Other alternative will be the small format aerial photography because of its acceptable geometric accuracy and its low costs. For regional scale mapping ASTER data can be also used which is easily available. Suitable models can be selected and applied, depending on availability of data. However, one should not forget about model calibration and validation. For this field data becomes very essential. 33 V. Case study The case study described here is on wind erosion feature mapping and monitoring. It is based on photo interpretation technique. In case that conventional aerial photos are too old or not available, small format aerial photography (SFAP) can be taken aboard a small aircraft and used to map and monitor erosion features. The case study is located in the central Rift valley of Kenya. It is based on the work of Hennemann and Nagelhout (2002). The case study consists of 2 parts: the analysis of the performance of SFAP and the analysis and assessment of wind erosion. A field study was carried out to map and monitor wind erosion features using small format aerial photography. The use of SFAP proved in spite of some minor operational drawbacks to be quite promising in terms of geometric accuracy, visual quality, cost effectiveness and above all timeliness. Generation and application of SFAP allowed for a detailed assessment of the current erosion status followed by a spatio-temporal GIS analysis of wind erosion patterns and trends in relation to key environmental factors. V.1 Description of case study Over the past two decades the semi-arid rangeland zone around Lake Naivasha in the central Rift Valley of Kenya has come under severe human pressure. Main causes are the steady encroachment into the area by smallholder farmers coming from higher parts of the Rift Valley, and the subsequent reduction of grazing land left for the Maasai pastoralists (Ataya, 2000). These developments have lead to overgrazing followed by severe wind erosion, which has now become a major threat to the livelihood of many inhabitants of the rangeland zone. Accurate analysis of wind erosion problems appeared to be impossible due to the overall lack of up-to-date, high-resolution remote sensing data of the affected area. This included the absence of recent aerial photographs, and although up-to-date satellite imagery (Landsat TM obtained in 1995 and 2000) was available for the rangeland zone, their general spatial resolution proved to be far too low for accurate mapping and monitoring of the prevailing wind erosion patterns. It 34 was therefore decided to apply small format aerial photography (SFAP) in order to achieve the required analysis. A special study was designed with as main objective to map and monitor wind erosion features using SFAP as a complementary tool to conventional aerial photography. Specific study objectives included (a) analysis of SFAP performance in terms of geometric accuracy and resource & cost effectiveness, (b) analysis and mapping wind erosion features, (c) assessment of current wind erosion status, (d) monitoring of wind erosion rate and (e) analysis of spatial relationships between current erosion status and trends with underlying erosion factors. V.2 Study area The study area is situated in the central part of the Kenya Rift Valley, southeast of Lake Naivasha, about 70 km northwest of Nairobi. The area lies at an altitude of around 2,100 m a.s.l. and covers about 370 ha. It is bounded by latitudes 0o 49' S to 0o 53' S and longitudes 36o 27' E to 36o 29'E and, administratively, falls under the Naivasha Division of Nakuru District. The area has a semi-arid to dry sub-humid, cool tropical highland climate. Mean annual rainfall ranges between 600-700 mm/year. Mean minimum and maximum monthly temperatures vary between 15.9oC to 18.5oC and from 24.6oC to 28.3oC, respectively. The long rainy season is from March to May with the short rainy season from October to December. Mean monthly wind velocity is highest in the period April - September (6-7 m/s) and lowest during November-February (3-4 m/s). Mean maximum wind speed is considerably higher and may reach up to 15-20 m/s during May-August resulting in high erosivity levels of the predominantly easterly winds during this period. The natural vegetation of the area mainly consists of low Acacia shrub grassland with Acacia drepanolobium ('Whistling Thorn') as main woody species and Themeda triandra as the dominant grass. Since the 1980s, however, most of the natural vegetation has been cleared or degraded into grassland. Current land use is mainly nomadic pastoralism with some marginal arable farming on small isolated farms, remnants of the smallholder settlement schemes that were abandoned in the early 1990s. 35 V.3 Methodology The generation of an adequate SFAP coverage of the study area requires careful planning. Figure 14 presents a flow-chart indicating the key parameters and successive steps required to determine the following : (a) number of flight lines, (b) number of photographs per flight line and (c) total number of SFAP photographs to be taken for a specified area, at the photo scale required. S u b j e c t / th e m e S tu d y a re a s iz e F o c a l l e n g th S c a le C am era m o d e l F lyin g h eig h t N e g a t iv e s iz e A irc ra ft g r o u n d s p e e d A r ea c o ve red p er p h o t o F o rw a rd o ve rl a p T im e in t e rv a l b e t w e e n ph otos O rie n t a t io n o f c a m e r a N e w a r e a p e r p h o to g r a p h N um ber of ph otos p e r fi lm N u m b e r o f p h o to s p e r fl i g h tl i n e L en g t h o f a fl ig h tli n e S id e w a rd o v e rl a p N um b e r o f f l ig h t li n e s T o ta l n u m b e r o f p h o t o s Figure 14: Flow-chart showing key parameters in planning (a) number of flight lines, (b) number of photographs per flight line and (c) total number of SFAP photographs for a specified area and required photo scale. To calculate above parameters, a simple spreadsheet in EXCEL was used. The following predefined equipment-related parameters were entered: focal length (35 mm), photo negative size (standard 24x36 mm) and ground speed (140 km/hour). Other parameters such as flying height, depression angle and forward overlap were defined rather by the objectives of the study 36 itself. Flying height, for example, is partly depending on photo-scale required. As the study of wind erosion features needs detailed photographs of scale of 1:5,000 on a 10x15cm photoformat, required flying height was determined at 2,300 feet above terrain. Further, stereoscopic analysis demands a forward overlap of at least 50%. From these predefined parameters the area covered per photograph, total photographs and time interval between photographs could be obtained using the above EXCEL spreadsheet. Five N-S parallel running flight lines were required to cover the reconnaissance area. Three coordinates of each flight line were entered in the GPS (Garmin III) of the aircraft. The first point, around 4 kilometres before entering the study area, was essential to position the aircraft so that it can enter the flight line in a straight way. To start and finish photographing, the first and the last point of the flight line above the reconnaissance area were also entered in the GPS. An important advantage of the Garmin III proved its capability to display the complete route (flight lines) including the position of the aircraft, so that during the flight small deviations from the flight course could be corrected. The SFAP photographs were taken with a Minolta X7000 AF camera (35mm) from a small Cessna 182 aircraft; a total of 130 photographs were shot covering the reconnaissance area following a carefully planned flight plan. From the above 130 photographs 28 photographs covering the actual pilot study area were finally selected for further analysis. A quality check was first done on the GPS-observations for ground control. During the field phase for each photograph around 20 GPS (Garmin 12) observations of 5 minutes were made to obtain ground control points to rectify the images. A photo-interpretation was then made of the pilot area using both old conventional aerial photographs of 1991 (1:20,000) and the new SFAP colour prints (1:5,000). During the field phase the prevailing soil pattern in the study area in combination with land cover and degradation features was systematically examined and mapped. This generated a geo-pedological map and two land cover/degradation maps, one for 1991 and one for 2000. To assess current status of land degradation in the area, the different erosion features were grouped into severity classes using an adapted version of the GLASOD methodology (Oldeman et al., 1991). From the erosion severity maps of the two different periods, general trends and rate of land degradation for 37 the period 1991-2000 could be calculated and assessed through standard GIS analysis using ILWIS. Finally, a spatio-temporal analysis was made of the main causal factors underlying the wind erosion patterns in the study area. V.4 Results and discussions 4.1 Analysis of SFAP performance 1) Geometric aspects The study showed that for a pilot area of 370 ha an up-to-date and adequate SFAP photo-cover at a scale 1:5,000 can be generated within a remarkably short time of about two weeks. The visual quality of the SFAP images ranged from good to excellent allowing for quick detection and detailed analysis of wind erosion features and other land degradation-related phenomena; height differences of less than 1 meter appeared to be readily detectable on the SFAP image. Main operational constraints included adverse weather conditions prior to flying (cloudiness) and during the flight (wind gusts), the latter in combination with the low speed as required and the low weight of the aircraft itself. GPS quality testing using established benchmarks yielded surprisingly good results with 85% of the observations within 6 m from the benchmark. Camera position analysis showed general inclinations between 0.1o and 5 o. When these values are matched with the accuracy limits used in conventional photography (< 3o), almost two-third of the SFAP images appears to fall within the acceptable range. Around 20 ground control points were obtained for each SFAP photograph. Each observation showed a distinct error but in the rectification process observations with considerable horizontal 38 errors were quickly detected as their root mean square error value (RMS value) is very high. Table 8 gives a general overview of the active control points and the corresponding sigma's in the rectification process. The overall sigma after rectification is low. Multiplying sigma value with pixel size gives errors ranging between 1.0 m up to 2.1 m, with a mean of 1.4 m per photograph. These values fall within the error range of the GPS (Garmin 12) recordings as indicated above. Table 8: Sigma values before and after correction, pixel size and calculated error per SFAP photo (projective transformation) SFAP number GPS points measured Sigma with all GPS points Active control points Sigma after correction Pixel size (m) Error (m) B00 14 7.58 9 2.36 0.41 0.97 B01 20 8.36 14 4.93 0.43 2.12 C18 19 4.80 14 3.26 0.50 1.63 C20 20 7.18 14 3.05 0.49 1.50 C22 24 5.79 17 2.78 0.48 1.34 Table 9: Number of man-days and costs (US$) for the various flight and fieldwork operations and materials to map an area of 370 ha SFAP operations and materials Number of Man-days Costs (US$) Flight preparation 1.0 Films & developing of photographs 1.0 60 Execution of flight 0.5 220 Collecting of GPS points (180) 3.0 Interpretation of photographs (14) and collecting wind erosion data 4.0 Scanning and georeferencing of photographs 1.5 Digitising and compilation of final map 2.0 Total resource use (370 ha) 2 2 Resource use efficiency in man-day / km and in US$ / km 13.0 280 3.5 76 39 2) Manpower and cost efficiency aspects Table 9 presents a general overview of the number of man-days and costs (US$) for the different flight and fieldwork operations to obtain a wind erosion status map of the study area (370 ha). In total, 13 working days were needed to cover an area of around 370 ha, 2.5 days for flight preparation and obtaining the photographs, 5 days of fieldwork and 5.5 days office work. However, this does not mean that the job was done in just two weeks: Between planning for the flight and the flight itself, almost two weeks were lost due to prohibitive (cloudy) weather conditions. In terms of delivery time SFAP compares quite favourably with conventional aerial photography which only in exceptional cases can deliver a similar output in such a short period. Also, overall costs for SFAP generation are minimal when compared with conventional aerial photography. For the study area (3.7 km2) and the total reconnaissance area of 15.3 km2 (both at scale 1:5,000) costs were US $ 76 and US$ 20 per km2, respectively. These values are only 25% and 7% of the photo costs (US$ 300/km2) incurred by an urban cadastral project in Bolivia using conventional photographs of a comparable scale (1:4,000). 4.2 Analysis and assessment of wind erosion 1) Soils and landscape The study area forms part of the central Rift Valley floor and consists largely of a gently sloping Volcanic Plain landscape covered with stratified volcanic ash deposits from the nearby Mt. Longonot volcano. The Volcanic Plain is almost entirely covered by young, poorly developed coarse-textured soils derived from Pleistocene Longonot ash and Akira pumice deposits. The soils generally consist of very deep, excessively drained, very friable, brown loamy sands and sands (Ah and Bw horizons) overlying a succession of dark grey and whitish grey, loose fine ash and pumice gravel layers (C-horizons). Three geopedological map units were distinguished on the basis of micro-relief consisting of low knolls (representing fossil dunes) : Pv111 - Volcanic Plain without knolly micro-relief, Pv112 - Volcanic Plain with low knolly micro-relief (0.5 to 2.0 m) and Pv113 - Volcanic Plain with moderate knolly micro-relief (2.0 to 4.0 m). Above soils classify as Areni-Vitric Andosols (Dystric) according to the World Reference Base classification (FAO 1998) on account of their sandy texture, relatively high content of volcanic glass in the 40 fine earth fraction, and low base saturation in the control section. The soils are subdivided on the basis of (a) thickness of the relatively coherent Bw horizon, and (b) presence and depth of loose, highly erodible sandy ash layers in the lower subsoil (C-horizon). Three soil phases with increasing vulnerability to wind erosion could thus be distinguished in the study area: Soil phase A : thickness Bw horizon. < 50 cm, immediately overlying pumice gravel layers Soil phase B : thickness Bw horizon > 50 cm immediately overlying loose fine ash layers Soil phase C : thickness Bw horizon < 50 cm immediately overlying loose fine ash layers 2) Classifying and mapping of wind erosion features The excellent visual quality and large photo-scale of the SFAP images allowed for a sound analysis and interpretation of the various wind erosion features. Detailed photo examination revealed the occurrence of a peculiar wind erosion pattern in the southern and central parts of the study area. This pattern comprises clusters of ENE-WSW running strips each consisting on its weather-side of a ‘head’, in the form of an oval to gully-shaped blow-out depression and a fanshaped depositional rear or ‘tail’ on its lee-side (Figure 15). The morphology of the blow-out depressions (wind erosion subtype D covering 0.5% of study area) is variable depending on their stage of development. Shape and size may range from oval, shallow depressions of only a metres long and wide at an early stage of development to massive deflation trenches of up to 150 m long and over 25 m wide in a mature stage. Depth of the blow-out depressions generally varies between 0.5 meter to 4.0 meter. Shallow blow-out depressions appear to be carved out into the Bw-horizon showing a brownish colour on the SFAP photo-image; Mature deflation trenches have cut through the lower subsoil into the underlying whitish to dark grey ash strata thus showing grey on the photo-image. The depositional areas or sand sheets (wind erosion subtype S covering 5.1% of study area) are mostly associated with one particular blow-out depression. The thickness of the sand sheets ranges from a few centimetres to over 50 cm. For the purpose of mapping both wind erosion subtypes D and S have been sub-divided into 7 different classes according to actual depth of deflation (abrasion) or deposition. 41 Figure 15: SFAP stereogram showing wind erosion pattern of coalescing deflation trenches (approximate scale 1: 3,400) (Hennemann and Nagelhout, 2002) 3) Assessment of current wind erosion status The assessment of current erosion status has been based on the concept of soil degradation severity as defined in the GLASOD classification (Oldeman et al., 1991). Soil degradation severity is an aggregation of 2 dimensions: (1) the degree of soil degradation (vertical dimension) and (2) the extent of the degradation process (lateral dimension). The degree or intensity of soil degradation is related to observed changes in the agricultural suitability, productivity, and restoration potential and biotic functions at one particular location(Oldeman and Lynden, 1997). The extent of soil degradation is defined as the relative frequency of occurrence of a particular type of degradation within the delineated map unit. The GLASOD classification has been applied in this study in adapted form to categorize and interpret wind 42 erosion features mapped. Four different erosion intensity classes were distinguished in both deflation and depositional areas; class distinctions and critical levels are defined as below. D - Blow-out depressions (deflation / abrasion areas) None: No visible signs of recent erosion. Slight: Topsoil is partly removed exposing brown Bw-horizon Moderate: Topsoil and upper subsoil have been completely removed and dark grey C-horizon is exposed deflation depth < 1 m. Severe: Topsoil and entire subsoil have been removed and C-horizon with light and dark grey ash-layers is exposed; deflation depth ranges between 1-4 m. S - Sand sheets (depositional areas) None: No visible signs of recent deposition. Slight: Depth of deposition is < 5 cm; some grasses are still present Moderate: Depth of deposition is 5 - 25 cm Severe: Depth of deposition > 25 cm Due to the high spatial resolution, large photo scale and good quality of the SFAP images in combination with the relative spatial homogeneity of the different wind erosion features themselves, it was possible to assess and spatially present the current wind erosion status directly on the basis of above erosion intensity classes (Figure 16). 43 Figure 16: Current wind erosion status of study (Hennemann and Nagelhout, 2002) However, analysis of the spatial relationship between wind erosion occurrence and the various geopedological units requires specific additional information about the relative extent of wind erosion features in a particular map unit. This is expressed as the frequency of occurrence within the delineated mapping unit. The following frequency classes have been used : Infrequent : < 5% of the map unit is affected; Common: 6 to 10% of the map unit is affected; Frequent : 11 to 25% of the map unit is affected ; Very frequent : 26 to 50% of the map unit is affected and Dominant : > 50% of the map unit is affected. To obtain wind erosion severity classes for the area, a 2D-matrix table was used combining erosion intensity and frequency classes into integrated erosion severity classes (Oldeman and Lynden, 1997). When applied to the three main geopedological map units in the study area, Pv111, Pv112 and Pv113, it appears that the latter unit is most severely affected (12.6 % in the high - very high severity class), followed by map unit Pv112 (4.6 % in the high - very high severity class) and finally, map unit Pv111 (0% in the high-very high severity class). 44 Table 10: Absolute and relative changes in area affected by wind erosion during 1991 - 2000 Year 1991 Area (ha) 2000 % study area Area (ha) Deflation areas (D) 0.22 0.1 1.89 Deposition areas (S) 1.14 0.3 19.00 % study area 0.5 5.1 4) Spatio-temporal analysis of wind erosion To monitor the impact of wind erosion on the area SFAP was used in combination with conventional photography to analyse surface changes over the period 1991 – 2000. They showed a nine-fold increase of deflation areas and a nearly twenty-fold increase of deposition areas (Table 10). In addition, it was found that gross and annual volumetric soil losses per ha in the area range between 53 – 64 m3/ha and 6-7 m3/ha/year, respectively. Assuming an average soil bulk density of 1.3 Mg/m3, total and annual soil weight losses from the area are estimated at 69-83 Mg/ha and 8-9 Mg/ha/year, respectively. For the most vulnerable map unit (Pv113) total and annual weight losses are quite considerable being 260 – 310 Mg/ha and 29-34 Mg/ha/year, respectively. Standard GIS analysis including map overlay revealed a strong spatial relationship between (a) severity of wind erosion and b) the occurrence of a pronounced knolly micro-relief in the study area. This is best exemplified by geopedological map unit Pv113 which combines a high erosion severity class with a distinct knolly micro-relief. It seems likely that the presence of knolly terrain forms a key factor in the wind erosion process as wind erodibility has been found to increase sharply in areas with a distinct micro-relief (Chepil et al., 1964). An additional causal factor is the predominance of wind erodible soil phase C in this map unit. Important parameters here are the poor aggregate size distribution (high % aggregates with diameter < 0.84 mm) in combination with a low specific density of the soil, probably due to the presence of pumice in the fine earth fraction. Both soil characteristics have a strong influence on wind erodibility (Zobeck, 1991). More on-site research is required on the specific relationship between wind erodibility and key mineralogical, pedological and soil physical properties of the different soils in the area. 45 Finally, additional analysis revealed that almost all large deflation trenches are found on or near abandoned arable fields. This confirmed earlier suspicions that careless soil management in combination with cattle trampling and destruction of old farm roads has been a major contributing factor to wind erosion in the area (Ataya, 2000). V.5 Conclusion The use of SFAP proves to be quite promising in terms of geometric accuracy, visual quality and spatial resolution in particular, manpower and cost efficiency and above all, timeliness. The study showed that for an area of 370 ha an up-to-date, adequate SFAP photo-cover at scale 1:5,000 can be generated within a remarkably short time (two weeks). Main operational constraints include adverse weather conditions prior to and during the flight, the latter in combination with the low speed and low weight of the aircraft itself. SFAP compares quite favourably with conventional aerial photography, particularly in terms of delivery time since the latter only in exceptional cases can deliver a comparable output within such limited periods. Finally, overall costs for SFAP generation are minimal in comparison to conventional aerial photography. Generation and application of SFAP allowed for a detailed assessment of the current erosion status in the study area followed by spatio-temporal analysis of wind erosion patterns and trends in relation to key environmental factors. References ACSAD, 2007. Desertification monitoring and assessment in the Arab World using satellite imageries between 1982 and 2005, The Arab Center for the Studies of Arid Zones and Dry Lands, Damascus. Allmaras, R.R., Burwell, R.E., Larson, W.E. and Holt, R.F., 1996. Total porosity and random roughness of interrow zone as influenced by tillage. USDA Conserv. Res. 7, US. Gov. Pritn. Office, Washington D. C. Ataya, C.O., 2000. Wind erosion of volcanic soils. A reconnaissance study in the Southern Catchment of Laka Naivasha Region, Kenya. MSc thesis Thesis, ITC, Enschede. Ayoub, A.T., 1998. Extent, severity and causative factors of land degradation in the Sudan. Journal of Arid Environments, 38(3): 397-409. Chepil, S., F. H. Siddoway, F.H. and al., e., 1964. Wind erodibility of knolly terrain. Journal of 46 soil and water conservation, 19 (5): 179-181. Chepil, W.S., 1942. Measurement of wind erosiveness of soils by the dry sieving procedure. Scientific Agriculture, 25: 154-160. Daniell, J.J. and Hughes, M., 2007. The morphology of barchan-shaped sand banks from western Torres Strait, northern Australia. Sedimentary Geology, 202(4): 638-652. de Jong, S.M., 1994. Applications of reflective remote sensing for land degradation studies in a mediterranean environment. Ph.D. Thesis, Utrecht University, Utrecht, The Netherlands, 237 pp. Eswaran, H., Lal, R. and Reich, P.F., 2001. Land degradation: An overview. In: E.M. Bridges et al. (Editors), Response to land degradation. Oxford & IBH Publishing Co. Pvt. Ltd., New Delhi, pp. 21-35. Goossens, D. and Riksen, M., 2004. Wind erosion and dust dynamics at the commencement of the 21st century. In: D. Goossens and M. Riksen (Editors), Wind erosion and dust dynamics: observations, simulation, modelling. ESW publications, Wageningen, pp. 713. Gregory, J.M., Wilson, G.R., Singh, U.B. and Darwish, M.M., 2004. TEAM: integrated, processbased wind-erosion model. Environmental Modelling & Software, 19(2): 205-215. Griffin, D.W., Garrison, V.H., Ferman, J.R. and Shinn, E.A., 2001. African desert dust in the Caribbean atmosphere: Microbiology and public health. Aerobiologia, 17: 203 - 213. Hennemann, G.R. and Nagelhout, A., 2002. Wind erosion mapping and monitoring in the central Rift valley of Kenya using small-format aerial photography (SFAP), ISCO XII, Beijing, China. Hennemann, G.R. and Nagelhout, A., 2004. Searching for effective, low-cost methods to detect and assess wind erosion damage: the promise of small-format aerial phtography (SFAP). In: D. Goossens and R. Michael (Editors), Wind erosion and dust dynamics: observations, simulations and modelling. ESW Publications, Wageningen, pp. 123-137. Jian, Z., F., Wang, L.K. and Peng, X., 1992. Wind erosion on a coastal hilly soil in China. In: H. Hurni and K. Tato (Editors), Erosion, conservation and small scale farming. Geographica Bernensia, Marceline, Missouri, pp. 121 - 131. Khresat, S.A., Rawajfih, Z. and Mohammad, M., 1998. Land degradation in north-western Jordan: causes and processes. Journal of Arid Environments, 39(4): 623-629. Klik, A., 2008. Wind Erosion Assessment in Austria Using Wind Erosion Equation and GIS. BOKU – University of Natural Resources and Applied Life Sciences Vienna. Lyles, L., 1983. Erosive wind energy distributions and climatic factors for the West. Journal of Soil and Water Conservation, 38(2): 106-109. Masri, Z., Zöbisch, M., Bruggeman, A., Hayek, P. and Kardous, M., 2003. Wind erosion in a marginal Mediterranean dryland area: a case study from the Khanasser Valley, Syria. Earth Surface Processes and Landforms, 28(11): 1211-1222. Oldeman, L.R., 1994. The global extent of land degradation. In: D.J. Greenland and I. Szabolcs (Editors), Land Resilience and Sustainable Land Use. CABI, Wallingford. Oldeman, L.R., Hakkeling, R.T.A. and Sombroek, W.G., 1991. World map of the status of human-induced soil degration: an explanaroty note. Gobal Assessment of Soil Degradation (GLASOD), second revised edition. ISRIC, Wageningen, The Netherlands. Oldeman, L.R. and Lynden, G.W.J., 1997. Revisiting the Glasod methodology. In: R. Lal, W.H. Blum, C. Valentine and B.A. Stewart (Editors), Methods for assessment of soil degradation. CRC Press, New York. 47 Panebianco, J.E. and Buschiazzo, D.E., 2008. Erosion predictions with the Wind Erosion Equation (WEQ) using different climatic factors. Land Degradation & Development, 19(1): 36-44. Renard, K.R., Foster, G.R., Weesies, G.A., McCool, D.K. and Yoder, D.C., 1997. Predicting soil erosion by water: a guide to conservation planning with the Revised Universal Soil Loss Equation (RÙSLE). Agriculture Handbook No. 703, U.S. Department of Agriculture. Shrestha, D.P., Soliman, A.S., Farshad, A. and Yadav, R.D., 2005. Salinity mapping using geopedologic and soil line approach, 26th Asian Conference on Remote Sensing, Hanoi, Vietnam, pp. 6. Smith, J.L. and Lee, K., 2003. Soil as a source of dust and implications for human health. Advances in agronomy, 80: 1-32. Tibke, G., 1988. Basic principles of wind erosion control. Agric. Ecosystem Environ.(22/23): 103-122. Tueller, P.T., Lent, P.C., Stager, R.D., Jacobsen, E.A. and Platou, K.A., 1988. Rangeland vegetation changes measure from helicopter-borne 35 mm aerial photography. Photogrammetric Engineering and Remote Sensing, 54: 609-614. UNEP and Leauge_of_Arab_States, 2004. State of desertification in the Arab World (Updated study), ACSAD, Damascus. Van der Knijf, J.M.f., Jones, R.J.A. and Montanarella, L., 1999. Soil Erosion Risk Assessment in Italy. Joint Research Centre, Space Applications Institute, European Soil Bureau, pp. 58. Wassif, M.M., Sharkawy, S.F., Genead, A.Y., Reda, M. and Atta, S.K., 2002. Evaluation of wind erosion models under Egyptian conditions. Egyptian Journal of Soil Science, 42(2): 219237 Williams, J.R., 1986. Effect of erosion productivity EPIC water erosion model. Proc. Fourth Fed. Interagency Sediment Conference, 6: 1-8. Williams, J.R. et al., 1990. EPIC - Erosion/Productivity Impact Calculator: 2. User Manual, Technical Bulletin No. 1768, U.S. Department of Agriculture. Woodruff, N.P. and Siddoway, F.H., 1965. A wind erosion equation. Proceedings of the Soil Science Society of America, 29: 602 - 608. Zhao, H.-L. et al., 2006. Wind erosion and sand accumulation effects on soil properties in Horqin Sandy Farmland, Inner Mongolia. CATENA, 65(1): 71-79. Zobeck, T.M., 1991. Soil properties affecting wind erosion. Journal of Soil and Water Conservation, 46(2): 112-118.