Spatial Modeling for Land Degradation Assessment Using Remotely Sensed Data and Geographic Information System; A Case Study of Daungnay Watershed, Magway District, Myanmar

GISdevelopment.net ---> Application ---> Environment

Spatial Modeling for Land Degradation Assessment Using Remotely Sensed Data and Geographic Information System; A Case Study of Daungnay Watershed, Magway District, Myanmar

Cung Chin Thang
Associate Expert (GIS/Natural Resources), International Center
for Integrated Mountain Development Nepal,

Dr.Apisit Euimnoh
Associate Professor(Rtd),
Asian Institute of Technology, Thailand,

Dr. Ganesh P.Shivakoti
Associate Professor, Asian Institute of Technology,
Thailand,

Dr. Roberto Clemente
Associate Professor, Asian Institute of Technology, Thailand

Abstract:
Land degradation in terms of soil erosion is a major environmental issue posing threat to sustainability livelihood in semi-arid region of Central Myanmar. This study attempted to assess land degradation severity by integrating biophysical and social parameters (population density) in spatial approach. Knowing extent and severity of the land is important as a decision support system to policy makers, resource managers as well as local communities and farmers. The result of this assessment showed spatial distribution of different land degradation severity across the area in watershed basis. Soil erosion susceptibility was estimated through Universal Soil Loss Equation (USLE). To examine land potential, land capability classification was carried out using Storie Index Rating (SIR). Population density was included in the model since human pressures on the land was considered as a major contribution factor for land degradation. Soil erosion, land capability and population density parameters were modeled in GIS platform. Though modeling of land degradation assessment is specific and involving much basic information such as suitability of crops, soil fertility, social and environment benefits etc., this study was only intend to develop a model based on spatial information and demographic data.

Introduction
Daungnay watershed is located in the heart of Dry Zone, in Magway District, Central Myanmar a distinct semi-arid region within the country, lying between N 20⁰ 10” to N to N 20⁰ 25” and E 94⁰ 55” to 95⁰ 15” with approximate 470 km² watershed area. Daungnay River is flowing from Northeast to Southwest and joint the Irrawaddy River at 8 km North of Magway town. The elevation ranges from almost flat to the peak of Setkyar Taung with a height of 375 m (1232 feet). The watershed is generally gentle to undulating (<10% of slope gradient) except Eastern part where a range of mountains locates as its boundary while the Northwest end of the watershed is rather steep due to the abrupt riverbank of Irrawaddy. Three townships administrative boundaries divide it into 12-village tracts covering 45 villages. Total population is 38,613 (1995) and mean population density is 82 inhab/km².

Rainy season normally started at mid-June with heavy rainfall and fallowed by a certain drought period and some rains at the end, meteorology department defined bimodal rainfall pattern. The minimum and maximum temperature is 90 C and 430 C respectively (District Meteorology Department, Magway).

Meadow Alluvial soils are found on low lands and near the main river. Lithosols on the West and Vertisols occupied most of the area. They derived from soft and unconsolidated sedimentary rock of young geologic age, most are sandy soils. As a result, moisture-holding capacity is generally low. In the topsoil, organic matter content is considerably low, resulting pale in color (Beerneart, 1995).

Figure 1. Location map of Daungnay watershed

Though Daungnay is a drought stream, the flow intensity during rain is considerably high. It flushed away the top soil layers and deposited to the lower land and at the Irrawaddy riverbank forming alluvial fans every year. Like other parts of semi arid regions, considerably low rainfall, usually less than 40 inches (100 mm) a year, natural vegetations is spare, poor and dominated by dry Scrub forests. The characteristic natural flora is dry savanna, with small thorny acacias and Euphorbia, cacti and short grasses.

Overall objective;
To develop a watershed based land degradation assessment model by geospatial analysis approach

Specific objectives;

to estimate the sediment produced by water erosion,
to classify land capability of the land based on Storie Index Rating (SIR),
to determine role of population density in land degradation

Methodology
The most important biophysical aspect for land use management could be understanding potential of the land and it can be represented by land capability and soil erosion susceptibility. Integrating distribution of population as possible human disturbances on the land, degradation of the land can be theoretically identifiable. This is the concept of geospatialized land degradation assessment used in this research.

Soil erosion susceptibility was determined by Universal Soil Loss Equation (USLE). To examine land potential, land capability classification was carried out using Storie Index Rating (SIR). As an element of land degradation, population density was classified and included in the model. Soil erosion, land capability and population density parameters were overlaid and modeled in GIS platform.

Soil Erosion Susceptibility
The most widely used method for predicting soil loss from overland areas is the Universal Soil Loss Equation (USLE) (Mitchel, J. K et al. 1980). The general form of the USLE, as expressed in metric units, is as follows (Goldman et al. 1986):

A = R*K*L*S*C*P

Where,

A = the average-annual soil loss

R = rainfall erosivity factor

K = soil erodibility factor

LS = slope length and steepness factor

C = vegetative cover factor

P = erosion control practice factor

Mean annual soil loss (A) has the same unit of K, which represent it in tons/ha per one unit of metric R.

Rainfall Erosivity (R)
Rainfall erosivity factor (R) is based on kinetic energy considerations of falling rain and represents a measure of the erosive force and intensity of rain in a normal year (Goldman et al. 1986). Two components of the factor are the total energy (E) and the maximum 30-min intensity of storms I30 (Wischmeier and Smith, 1978). The R-factor is the sum of the product of these two components for all major storms in the area during an average year. Even though EI30 is the most reliable source for computing R, other equations might be used where E and I30 were not available. R = 38.5 + 0.35 P (P = mean annual precipitation) gives acceptable erosivity index for tropical and subtropical ecological zones (Eiumnoh, A., 2000). In the study area, there were no rainfall stations at that time and R-value was interpolated through annual rainfall data (1995) from surrounding stations instead. Resultant R- value ranged from 301 to 406.

Soil Erodibility Index (K)
Soil erodibility index (K) is influenced by soil structure, organic matter content, soil texture and soil permeability, it should be based on measured value wherever possible (Morgan, 1995). The K value can be calculated with the use of nomograph, derived by Wischemier and Smith (1978), when all the values of K influencing factors are available.

Figure 2. Land degradation assessment methodology

Lack of detailed soil map forced to use data from soil sampling analysis. Eight samples were taken evenly distributed over existing soil series across the watershed. From the analysis report, all the soil types in the study area showed strong sand component, up to 70%. Sands being non-coherent and structureless are easily erodible, high specific gravity, low total porosity, low available water range, and high infiltration and course textured soil (FAO, 1975). Applying nomograph, K- values ranged from 0.28 to 0.55 in the study area.

Topographic Factor (LS)
The factors of slope length (L) and slope steepness (S) are combined in a single topographic index termed LS factor. Wischmeier and Smith defined the slope length as the distance from the point of origin of overland flow to the point where either the slope gradient decreases enough that deposition begins or the runoff water enters a well-defined channel that may be part of a drainage network or a constructed channel. Slope and aspect maps were generated from digital elevation model (DEM) formed from contour map of 50 feet (15.24 m) interval. The following equations were applied to produce LS.

LS - factors were computed in TNTmips 6.5 software since it has ability to run programs using Simple Macro Language (SML) script.

Crop Management factor (C)
The ratio of soil loss under given crop to that from bare soil is represented as crop management factor (C). In order to determine C factor, Daungnay watershed was classified into 5 land uses generated from Landsat TM images (1995), Row-133 and 134, path 46, applying maximum likelihood of supervise classification. C- values were defined 0.06 for natural woodland (open), 0.014 for scrub and grass lands, 0.01 for built-up area, 1 for bare land and 0.377 for agriculture land as recommended by Morgan, 1995.

Erosion Control Practice Factor (P)
This factor is a ratio between erosion occurring in a field treated with conservation measures and another reference plot without treatment. Therefore, erosion control practice factor is based on the soil conservation practices operated in a particular area. In this study, P value for non- agricultural lands is assigned as 1 and 0.5 was assign for cultivation areas having erosion control practices.

Land Capability Classification
Land capability classification was executed to determine land potential as well as relationship between soil erosion and land capability classes. Even though availability of different land capability classification systems, Storie Index Rating (SIR) developed by Storie, R.E. (1933 and 1978) was chosen for its simplicity. The SIR system is an index for numerical rating of soils and expresses numerically the relative degree of suitability, or value of a soil for general land use and agriculture. The rating is based on soil and topographic characteristics only and obtained by evaluating specific soil factors. Other factors such as availability of water for irrigation and climate are not considered. Four general factors are considered in the Storie index rating (SIR) as follows;

Population Density
Accelerated erosion and excessive runoff are connected with development activities and human disturbances: clearance of fragile zones, denudation and compaction of soil through overgrazing, exhaustion of soil through intensive cropping. Erosion increases as a function of population density in a given agrarian system, if the population passes a certain threshold, land starts to run short, and soil restoration mechanisms seize up (Pieri, 1989). In Sudano-Sahelian zones, when the population exceeds 20-40 inhabitants/km², the fallow period is shortened to the point of ineffectiveness, and one speaks of a densely populated degraded area when the population reaches 100 inhab/km² (FAO, 1996).

Dry zone, central Myanmar, being semi-arid region, assuming similar trend as in Sahel, population density was classified in to 3 classes, < 50 inhab/km2 as sparse, 50- 100 inhab/km2 as moderate and >100 inhab/ km2 as dense population. Finally, Land degradation extent was determined using spatial overlay function with weighing system in GIS.

Results and Discussion
Soil Erosion Susceptibility
From USLE analysis, more than 60% of total area covered by forest and agriculture land with gentle slopes has low soil loss, 1-10 tons/ha/year while moderate erosion rates were found at bare lands and on steep boulders accounting 33% of the total land area. High soil loss can be seen in the steep east and west end of the watershed. In Daungnay watershed, about one third is under critical condition as if annual soil loss of more than 12.5 ton/ha/yr is considered as critical value in South East Asia and it is rather large area compare with the watershed as a whole.

Land Capability Classes
The index rating for a soil is obtained by multiplying the factors, A, B, C, and X and dividing by 100, thus each factor contributes to the final rating. A rating of 100 per cent expresses the most favorable or ideal condition (class 1), and lower percentage ratings are given for conditions that are less favorable to crop production (class 8). There is no class 1 in Daungnay watershed and class 2 accounts for 70 %, class 3 for 26.13 % and the rest for 3%.

Figure 3. Resultant maps of land degradation assessment

Population Density Map
Population Density map was prepared in village track wise and classified into 3 groups. Dense populations were found where vast arable land with good accessibility. There are 4 village tracks with population density of more than 100/km2 located near to the highway and Irrawaddy River. The lowest population can be seen village tracks of steep slope occupying forest and limited cultivation land while moderate density is in large cultivation land and gentle slope.

Land Degradation Assessment
The three spatial datasets were prepared separately according to their respective categories to develop land degradation severity information by spatially modeling approach. Since rainfall induced soil erosion was the most influenced factor in land degradation, more weight was assigned accordingly. Despite of the fact that land capability classes might not reveal a strong relationship to soil erosion susceptibility, it was considered important role in the context of land degradation assessment. Similarly, there was a function of risk of erosion by increase of population density as well (FAO, 1996).

Applying simple arithmetic calculation in ARCVIEW GIS software, final land degradation extent was produced.

Figure 4. Land degradation extent in Daungnay Watershed

Area Correlation Analysis
Soil Loss and Land use/land cover In general, any kind of vegetation coverages reduced the splash and runoff erosion. This study also agreed with this phenomenon. Soil loss from all forested area was the lowest i.e., 0-10 ton/ha/yr, among all land use classes. Agricultural land contributed the most soil loss followed by thrones and shrubs land.

Soil erosion and population density
Generally, soil erosions produced from different population density classes were similar, 30.61%, 35.77% and 33.57% from less, moderate and high population density respectively. But there was a slight significant contribution by middle class (50-100 inhab/km2) than the other two.

Soil loss and land capability
Both were largely influenced by soil properties. Despite the fact that higher capability classes (1, 2, 3) might not be considered less susceptibility to soil erosion for their better quality, no erosion and slight erosion were occurred at high capability class in this case.

Land Degradation and Soil Erosion
It was found that the estimated soil loss in 1995 by USLE was directly related to land degradation severity assessment. Soil erosion rate of 10-50 ton/ha/yr was account for the most severe land degradation since almost all of combined very severe and severe classes (15.02 % out of 17.69%) was fall into 10-50 ton/ha/yr annual soil loss class.

Land Degradation and Land Capability
Most of slight degradation problems were found in good land capability class as in 34.18 % out of total slight degradation of 39.75% was found in land capability class 2.

Land Degradation and Population Density
It is obvious that increase of population worsened erosion type of land degradation particularly in places where agriculture is a sole economy. In Daungnay watershed, total area of very severe and severe degradation is 17.89 % of the whole watershed and more than two-third was found in highest populated area (>100 inhab/km2). Inversely, more than half of total slight degradation areas were found where sparse population density (<50inhab/km2), 21.31% out of total 39.75%.

Conclusions
In this study, land degradation severity was determined by modeling geospatial layers of the most concerned parameters within remote sensing and GIS capabilities. Effects from water induced erosion, differences in land quality and population pressure have inevitably worsened depletion of the prone dry zone ecosystem of Central Myanmar. Attempt has been made to identify those degraded areas in different level of severity formed by effects from rain, soil and human pressure in selected area having 474.55 km².

Computed soil loss, is ranging from 0 to more than 200 ton/ha/yr. However, annual soil loss from more than 60% of total area is between 0- 10 ton/ha/yr and more than 200 ton/ha/yr is negligible (0.14 %).
Up to 30 % was found under critical condition as if annual soil loss of more than 12.5 ton/ha/yr is considered as critical value In South East Asia.
Although stand condition was poor, annual soil from forested area was still the lowest among all land use classes.
Land capability classification based on SIR implied its applicability that had testified though the classes might not reflect the actual land capability since limited soil information had to be applied in this SIR analysis.
This model proved that there was a strong relationship between population density and land degradation after combined with erosion susceptibility and land capability even though soil erosion and population density relationship was not much significant.

Recommendations
It needed to utilize higher number of soil samples, representing each soil groups of the study area for a more reliable soil erodibility and land capability classification. Assigning weights and scores to factors affecting land degradation should carefully develop to reflect actual condition. Though this study did not consider socioeconomic data except population, aspect of people awareness on environmental degradation should consider being a crucial factor for deterioration of land.

References:

ADB, 1991. Remote Sensing and Geographic Information System for Natural Resources Management, ADB environmental paper No.9. Manila Philippines.
Balba, A.M., 1995. Management of problem soils in Arid Ecosystems, CRC press, Inc., Florida.
Bancy, M.M et al., 2000. Assessment of Erosion Hazard with USLE and GIS: the Upper Ewaso Ng’iro North basin of Kenya, International Journal of Applied Earth Observation and Geoinformation, 2:78-86.
Berneart, F.R., 1995. Natural Resources Inventory in The Townships of Chaung U, Kyaukoadaung and Magway (Dry Zone of Myanmar), Consultant’s Report, Agriculture Development and Rehabilitation Development in the Dry Zone Project (MYA/93/004), FAO, Yangon, Myanmar.
Burrough, P.A. 1986. Principles of Geographic information Systems for Land Resources Assessment. Clarendon Press, Oxford, United Kingdom.
Burrough, P.A & Mcdonnell R.A., 1998. Principal of Geographic Information Systems, Oxford University Press, New York.
Cassettari .S., 1993. Introduction to Geo-information Management, Chapman & Hall, London.
Cambell, J.B. 1987. Introduction to Remote Sensing. Guidford, Newyork.
Carucci, V.F.P., 2001. Guidelines on Soil and Water Conservation for the Myanmar Dry Zone (Revised Edition), UNDP, Yangon.
Dwivedi, R.S. et al., 1997. The Inventory and Monitoring of Eroded Lands using Remote Sensing Data. International Journal of Remote Sensing, Vol. 18, No. 1, 107- 119.
Eiumnoh, A., 2000. Integration of Geographic Information Systems (GIS) and Satellite Remote Sensing (SRS) for Watershed Management, Technical Bulletin 150.Food & Fertilizer Technology Center, Taiwan.
FAO, 1973. Calcareous Soils, SOILS BULLETIN 21, Rome.
FAO, 1975. Sandy Soils, SOILS BULLETIN 25. Report of the FAO/UNDP Seminar on Reclamation and Management of Sandy Soils in the Near East and North Africa, Rome, Italy.
FAO, 1976. A Frame work for Land Evaluation, Soil resources development and conservation service, land and water development division, soil bulletin No.32, Rome, Italy.
FAO, 1987. Soil and Water Conservation in Semiarid Areas, FAO Soil Bulletin No. 57, Rome, Italy.
FAO, 1990. Erosion Mapping Using High-resolution Satellite Data and Geographic Information System. Pilot Study in State of Parana, FAO RSC Series No. 56, High Resolution Satellite Data Series No. 2,Rome, Italy.
FAO, 1996.Land husbandry - Components and strategy. Soils Bulletin No. 70, Rome. Fuhu, R., 1995. GIS for Natural Resources Management. Remote Sensing for Tropical Ecosystem Management, ESCAP.
Goldman, S. J., K. Jackson, and T. A. Bursztynsky. 1986. Erosion and Sediment Control Handbook. McGraw-Hill, New York.
Hudson, N., 1971. Soil Conservation, Cornell University Press, New York. ILRI, 1994. Drainage Principles and applications (Second Edition), International Institute for Land Reclamation and Improvement, Wageningen, The Netherlands.
Kywe Myint, U., 1998. Forest Management Plan, Magway District (1996-97 to 2005- 2006), Vol.1, Forest Department, Ministry of Forestry, Myanmar.
Lakshmana, R.K.M., 1995. Remote Sensing for land use planning, International Journal of Remote Sensing, Vol. 16, No. 1, 53-60.
Lal, P., 1988. Soil Erosion Research Methods. International Society of Soil Science, Wageningen, The Netherlands.
Lal .R et al., 1998. Assessment of Soil Degradation, CRC press, Florida, United States.
Lillesand. T.M and R.W Keifer. 2000. Remote Sensing and Image Interpretation. (Fourth Edition). John Wiley and Sons, Singapore.
Mitchel, J. K., and G. D. Bubenzer. 1980. "Soil Loss Estimation." In Soil Erosion, eds. M.J. Kirby and R.P.C. Morgan. Wiley, New York.
Morgan, R.P.C., 1995. Soil Erosion and Conservation (Second Edition). Longman Group, UK Limited.
Pahari, K.J., 1993. Soil Erosion Susceptibility using Remote Sensing and GIS: A Case Study of Andhikhola Watershed, Nepal, Master Thesis, Asian Institute of Technology, Bangkok, Thailand.
Pieri, C. 1989. Fertilité des terres de Savannes. CIRAD-Institute, Recherches Agronomiques Tropicales, Paris.
Rathanayake, W.M.U.K, 1999. Application of GIS for Soil Erosion Control Planning: A Case Study of Kandy District, Sri Lanka. AIT, Bangkok.
Sah, B.P., Honda, K., and Murai, S., 4997. Land Degradation and Socio-Economic Status Modeling, using Remote Sensing and GIS, for Watershed Management, Asian – Pacific Remote Sensing and GIS Journal Vol 9, No 2: 51-60.
Schwab, et. al., 1993. Soil and Water Conservation Engineering. United States: John Willey & Sons, Inc.
Storie, R.E. 1978. Storie index rating. University of California, Division of Agricultural Science Special Publication 3203.
Wilkie, D.S and Finn, J.T., 1996. Remote Sensing Imagery for Natural Resources Monitoring, A Guide for First-time Users, Columbia University Press, New York.
Wischmeier, W.H and Smith, D.D., 1978. Predicting Rainfall Erosion Losses – A Guide to Conservation Planning. Agriculture Handbook No. 537. United States.
WRI, 1999.DRYLANDS POPULATION ASSESSMENT II, World Resources Institute, UNDP Office to Combat Drought and Desertification, Washinton.
Zin, M.Thant., 1998. A Simulation Technique for Resources Management: A Case Study of Mount Popa, Myammar, Master Thesis, Kasetsart University, Bangkok, Thailand.