Advanced search

Sciences in Cold and Arid Regions ›› 2022, Vol. 14 ›› Issue (1): 32-42.doi: 10.3724/SP.J.1226.2022.21003.

Previous Articles    

Desertification status mapping in Muttuma Watershed by using Random Forest Model

S. Dharumarajan1(),Thomas F. A. Bishop2   

  1. 1.ICAR-National Bureau of Soil Survey and Land Use Planning, Regional Centre, Hebbal, Bangalor 560024, India
    2.Faculty of Agriculture and Environment, The University of Sydney, Sydney, NSW 2006, Australia
  • Received:2021-01-08 Accepted:2021-10-11 Online:2022-02-28 Published:2022-03-03
  • Contact: S. Dharumarajan E-mail:sdharmag@gmail.com

RichHTML

49

PDF (PC)

298

Abstract:

Potential of the Random Forest Model on mapping of different desertification processes was studied in Muttuma watershed of mid-Murrumbidgee river region of New South Wales, Australia. Desertification vulnerability index was developed using climate, terrain, vegetation, soil and land quality indices to identify environmentally sensitive areas for desertification. Random Forest Model (RFM) was used to predict the different desertification processes such as soil erosion, salinization and waterlogging in the watershed and the information needed to train classification algorithms was obtained from satellite imagery interpretation and ground truth data. Climatic factors (evaporation, rainfall, temperature), terrain factors (aspect, slope, slope length, steepness, and wetness index), soil properties (pH, organic carbon, clay and sand content) and vulnerability indices were used as an explanatory variable. Classification accuracy and kappa index were calculated for training and testing datasets. We recorded an overall accuracy rate of 87.7% and 72.1% for training and testing sites, respectively. We found larger discrepancies between overall accuracy rate and kappa index for testing datasets (72.2% and 27.5%, respectively) suggesting that all the classes are not predicted well. The prediction of soil erosion and no desertification process was good and poor for salinization and water-logging process. Overall, the results observed give a new idea of using the knowledge of desertification process in training areas that can be used to predict the desertification processes at unvisited areas.

Key words: desertification processes, vulnerability indices, Random Forest Model, extrapolation

Figure 1

Location of (a) the study area in Australia (b) Muttuma Watershed"

Table 1

Assigned scores of soil quality indices for calculation of DVI"

ClassScoreClassScore
Soil depthDrainage
Deep >100 cm1.0Well drained1.0
Moderate 50-1001.5Imperfectly drained1.5
Shallow <502.0Poorly drained2.0
Soil forming processesRunoff
Erosional2.0Slight1.0
Colluvial1.6Moderate1.5
Transferral1.2Severe2.0
Alluvial1.0Soil Quality Indices
Soil textureGood1.00-1.33
Coarse loamy, fine-loamy1.0Moderate1.33-1.66
Fine, very fine1.5Poor1.66-2.00
Coarse2.0
Organic Carbon
<0.5%2.0
0.5%-1%1.5
>1%1.0

Table 2

Assigned scores of terrain parameters for calculation of DVI"

AspectScoreSlopeScore
N1.00-31.00
NE, NW1.13-81.33
W1.28-151.66
E1.415-252.00
SE, SW1.9
S2.0

Table 3

Assigned scores of vegetation quality parameters"

ClassScoreClassScore
Fire riskPlant cover
Annual crops, modified grazing lands1.00High (>40%)1.00
Permanent grasslands, perennial pastures, perennial horticulture1.50Moderate (10%-40%)1.80
Woodlands, native forest2.00Low (<10%)2.00
Erosion protectionScoreResistance to droughtScore
Woodlands, native forest1.00Woodlands, native forest1.00
Permanent grassland, perennial pastures, perennial horticulture1.33Permanent grassland, perennial pastures, perennial horticulture1.33
Intensive animal husbandry, modified grazing lands1.66Intensive animal husbandry, modified grazing lands, Annual crops associated with permanent crops1.66
Dryland annual crops, non irrigated arable land2.00Dry land annual crops2.00

Table 4

Assigned scores of land utilization index"

Land useScoreLCCScore
Annual crops, modified grazing lands1.0I&II2.00
Permanent grasslands, perennial pastures, perennial horticulture1.5III&IV1.50
Woodlands, native forest2.0>V1.00

Table 5

Description of desertification vulnerability indices classes"

DVI classRangeDescription
Nil to slight1.00 ≤ DVI < 1.25Slight or non-threatened areas to desertification subjected to very low or no erosion with no or slight salinization risk
Potential1.25 ≤ DVI < 1.50Potential areas to desertification subjected to low to moderate erosion rates or low salinization risk
Fragile1.50 ≤ DVI < 1.75Fragile areas to desertification subjected to moderate to high erosion rates due to intensive cultivation or overgrazing or frequent fires; or subjected to moderate salinization risk
Critical1.75 ≤ DVI ≤ 2.00Critical areas to desertification and subjected to very high erosion rates due to intensive cultivation, overgrazing, frequent fires; or subject to high or very high salinization risk

Table 6

Datasets used for model building"

PredictorSourceSpatial resolutionTypeRange
Temperature (max)Bureau of Meteorology-Q19.7-22.8
Rainfall (mm)Bureau of Meteorology-Q556-739
Evaporation (mm)Bureau of Meteorology-Q1,424-1,557
SlopeASTER GDEM30 mQ0.0-0.45
AspectASTER GDEM30 mQ0.0-6.28
LS factorASTER GDEM30 mQ0.0-9.39
Wetness indexASTER GDEM30 mQ4.4-24.9
Sand 0%-30% (cm)Soil map grid of Australia3 secQ30.7-63.9
pH (0%-30 (cm))4.4-5.7
Clay (0%-30% (cm))10.3-44.2
Sand (60%-100% (cm))Soil map grid of Australia3 secQ6.5-44.5
pH (60%-100% (cm))3.6-5.5
Clay (60%-100% (cm))2.5-35.1
Organic carbonSoil map grid of Australia3 secQ0.49-1.91
NDVIMOD13Q1 (2000-2015)250 m, 16 daysQ0.34-0.67
EVIMOD13Q1 (2000-2015)250 m, 16 daysQ0.19-0.44
Vulnerability indicesClimatic quality index-C-
Terrain quality Index-C-
Vegetation quality index--
Land Utilization Index-C-
Soil Quality index-C-
DVI-C-

Figure 2

Training areas for desertification prediction"

Figure 3

Desertification vulnerability index map"

Table 7

Confusion matrix for training (pixels) and testing (pixels)"

ClassTraining dataTesting data
EWLNDSTotalEWLNDSTotal
Erosion (E)1,337347311,8144,68683,723418,458
Waterlogging (WL)1212138435576934666488
No desertification (ND)150264,50864,6907,62421728,45862836,927
Salinization (S)23545611510219763263
Total1,4902445173676,97412,41929532,62479846,136

Figure 4

Predicted desertification map"

Figure 5

Random forest importance of covariates on desertification processes"

Adhikari K , Kheir RB , Greve MB , et al. , 2013. High-resolution 3-D mapping of soil texture in Denmark. Soil Science Society of America Journal, 77: 860-876. DOI: 10.2136/sssaj2012.0275 .
doi: 10.2136/sssaj2012.0275
Akpa SIC , Odeh IOA , Bishop TFA , et al. , 2014. Digital mapping of soil particle-size fractions for Nigeria. Soil Science Society of America Journal, 78: 1953-1966. DOI: 10.2136/sssaj2014.05.0202 .
doi: 10.2136/sssaj2014.05.0202
Bakr N , Weindorf DC , Bahnassy MH , et al. , 2012. Multi-temporal assessment of land sensitivity to desertification in a fragile agro-ecosystem: environmental indicators. Ecological Indicators, 15(1): 271-280. DOI: 10.1016/j.ecolind. 2011.09.034 .
doi: 10.1016/j.ecolind. 2011.09.034
Basso F , Bove E , Dumontet S , et al. , 2000. Evaluating environmental sensitivity at the basin scale through the use of geographic information systems and remote sensed data: an example covering the Agri basin (Southern Italy). Catena, 40: 19-35. DOI: 10.1016/S0341-8162(99)00062-4
doi: 10.1016/S0341-8162(99)00062-4
Breiman L , 2001. Random forests. Machine Learning, 45: 5-32. DOI: 10.1023/A:1010933404324 .
doi: 10.1023/A:1010933404324
Bureau of Meteorology , 2011. Climate Statistics for Australian Locations-Cootamundra. .
Choubin B , Mosavi A , Alamdarloo EH , et al. , 2019. Earth fissure hazard prediction using machine learning models. Environmental Research, 179: 108770. DOI: 10.1016/j.envres. 2019.108770 .
doi: 10.1016/j.envres. 2019.108770
Cutler DR , Edwards Jr TC , Beard KH , et al. , 2007. Random forests for classification in ecology. Ecology, 88: 2783-2792. DOI: 10.1890/07-0539.1 .
doi: 10.1890/07-0539.1
Dasgupta A , Sastry KLN , Dhinwa PS , et al. , 2013. Identifying desertification risk areas using fuzzy membership and geospatial technique -A case study, Kota District, Rajasthan. Journal of Earth System Science, 122: 1107-1124. DOI: 10.1007/s12040-013-0331-x .
doi: 10.1007/s12040-013-0331-x
Dharumarajan S , Lalitha M , Natarajan A , et al. , 2017. Biophysical and socio-economic causes for increasing fallow land in Tamil Nadu. Soil Use Management, 33: 487-498. DOI: 10.1111/sum.12361 .
doi: 10.1111/sum.12361
Dharumarajan S , Bishop TFA , Hegde R , et al. , 2018. Desertification vulnerability index—an effective approach to assess desertification processes: A case study in Anantapur District, Andhra Pradesh, India. Land Degradation and Development, 29: 150-161. DOI: 10.1002/ldr.2850 .
doi: 10.1002/ldr.2850
Dharumarajan S , Hegde R , Janani N , et al. , 2019. The need for digital soil mapping in India. Geoderma Regional, 16. DOI: 10.1016/j.geodrs.2019.e00204 .
doi: 10.1016/j.geodrs.2019.e00204
Dharumarajan S , Hegde R , 2020. Digital mapping of soil texture classes using Random Forest classification algorithm. Soil Use Management, 1-15. DOI: 10.1111/sum.12668 .
doi: 10.1111/sum.12668
Dharumarajan S , Lalitha M , Parmar M , et al. , 2020. Desertification and land degradation vulnerability assessment using integrated geospatial techniques. Indian Journal of Soil Conservation, 48: 166-173.
Dou J , Yamagishi H , Pourghasemi HR , et al. , 2015. An integrated artificial neural network model for the landslide susceptibility assessment of Osado Island, Japan. Natural Hazards, 78: 1749-1776. DOI: 10.1007/s11069-015-1799-2 .
doi: 10.1007/s11069-015-1799-2
Friedl MA , Woodcock C , Gopal S , et al. , 2000. A note on procedures used for accuracy assessment in land cover maps derived from AVHRR data. International Journal of Remote Sensing, 21: 107-31077. DOI: 10.1080/014311600210434 .
doi: 10.1080/014311600210434
Grinand C , Arrouays D , Laroche B , et al. , 2008. Extrapolating regional soil landscapes from an existing soil map: Sampling intensity, validation procedures, and integration of spatial context. Geoderma, 143: 180-190. DOI: 10.1016/j.geoderma.2007.11.004 .
doi: 10.1016/j.geoderma.2007.11.004
Grundy MJ , Viscarra Rossel RA , Searle RD , et al. , 2015. Soil and landscape grid of Australia. Soil Research, 53: 835-844. DOI: 10.1071/SR15191 .
doi: 10.1071/SR15191
Hong H , Pourghasemi HR , Pourtaghi ZS , 2016. Landslide susceptibility assessment in Lianhua County (China): A comparison between a random forest data mining technique and bivariate and multivariate statistical models. Geomorphology, 259: 105-118. DOI: 10.1016/j.geomorph.2016. 02.012 .
doi: 10.1016/j.geomorph.2016. 02.012
Imbrenda V , D'emilio M , Lanfredi M , et al. , 2014. Indicators for the estimation of vulnerability to land degradation derived from soil compaction and vegetation cover. European Journal of Soil Science, 65: 907-923. DOI: 10.1111/ejss. 12184 .
doi: 10.1111/ejss. 12184
Keesstra S , Mol G , De Leeuw J , et al. , 2018. Soil-related sustainable development goals: Four concepts to make land degradation neutrality and restoration work. Land, 7(4):133. DOI: 10.3390/land7040133 .
doi: 10.3390/land7040133
Kidd DB , Malone BP , McBratney AB , et al. , 2014. Digital mapping of a soil drainage index for irrigated enterprise suitability in Tasmania, Australia. Soil Research, 52: 107-119. DOI: 10.1071/SR13100 .
doi: 10.1071/SR13100
Kosmas C , Ferrara A , Briasouli H , et al. , 1999. Methodology for mapping Environmentally Sensitive Areas (ESAs) to Desertification. In: the Medalus Project: Mediterranean Desertification and Land Use Manual on Key Indicators of Desertification and Mapping Environmentally Sensitive Areas to Desertification. In: Kosmas C, Kirkby M, Geeson N (eds.), European Union 1882, ISBN 92-828-6349-2, pp: 31-47.
Ludwig JA , Tongway JD , 1995. Desertification in Australia: an eye to grass roots and landscapes. Environmental Monitoring and Assessment, 37: 231-237. DOI: 10.1007/BF00546891 .
doi: 10.1007/BF00546891
McBratney AB , Mendonça ML , Minasny B , 2003. On digital Soil Mapping. Geoderma, 117: 3-52. DOI: 10.1016/S0016-7061(03)00223-4 .
doi: 10.1016/S0016-7061(03)00223-4
Moran JM , Bui EN , 2002. Spatial data mining for enhanced soil map modelling. International Journal of Geographical Information Science, 16: 533-549.
Mosavi A , Sajedi-Hosseini F , Choubin B , et al. , 2020a. Susceptibility mapping of soil water erosion using machine learning models. Water, 12(7): 1995. DOI: 10.3390/w12071995 .
doi: 10.3390/w12071995
Mosavi A , Golshan M , Janizadeh S , et al. , 2020b. Ensemble models of GLM, FDA, MARS, and RF for flood and erosion susceptibility mapping: a priority assessment of sub-basins. Geocarto International, 1-20. DOI: 10.1080/10106049. 2020.1829101 .
doi: 10.1080/10106049. 2020.1829101
Mosavi A , Sajedi HF , et al. , 2021. Susceptibility mapping of groundwater salinity using machine learning models. Environmental Science and Pollution Research, 28(9): 10804-10817. DOI: 10.1007/s11356-020-11319-5 .
doi: 10.1007/s11356-020-11319-5
Muchoney DM , Strahler AH , 2002. Pixel-and site-based calibration and validation methods for evaluating supervised classification of remotely sensed data. Remote Sensing and Environment, 84: 290-299. DOI: 10.1016/S0034-4257(02)00006-8 .
doi: 10.1016/S0034-4257(02)00006-8
Odeh IOA , McBratney AB , 2000. Using AVHRR images for spatial prediction of clay content in the lower Namoi valley of eastern Australia. Geoderma, 97: 237-254. DOI: 10. 1016/S0016-7061(00)00041-0 .
doi: 10. 1016/S0016-7061(00)00041-0
Pickup G , 1998. Desertification and climate change-the Australian perspective. Climate Research, 11: 51-63. DOI: 10.3354/cr011051 .
doi: 10.3354/cr011051
Prasad AM , Iverson LR , Liaw A , 2006. Newer classification and regression tree techniques: Bagging and random forests for ecological prediction. Ecosystems, 9: 181-199. DOI: 10.1007/s10021-005-0054-1 .
doi: 10.1007/s10021-005-0054-1
Rajbanshi J , Das S , 2021. Monitoring land sensitivity to desertification using the ESAI approach and evaluation of the key indicators: A spatio-temporal study in India. Land Degradation and Development, 32: 3045 -3061. DOI: 10.1002/ldr.3965 .
doi: 10.1002/ldr.3965
Rogan J , Miller J , Stow D , et al. , 2003. Land-cover change monitoring with classification trees using landsat TM and ancillary data. Photogrammetric Engineering and Remote Sensing, 69: 784-793. DOI: 10.14358/pers.69.7.793 .
doi: 10.14358/pers.69.7.793
Salvati L , Zitti M , 2008. Assessing the impact of ecological and economic factors on land degradation vulnerability through multiway analysis. Ecological Indicators, 9: 357-363. DOI: 10.1016/j.ecolind.2008.04.001 .
doi: 10.1016/j.ecolind.2008.04.001
Kapović Solomun M , Barger N , Cerda A , et al. , 2018. Assessing land condition as a first step to achieving land degradation neutrality: A case study of the Republic of Srpska. Environmental Science and Policy, 90: 19-27. DOI: 10.1016/j.envsci.2018.09.014 .
doi: 10.1016/j.envsci.2018.09.014
Symeonakis E , Karathanasis N , Koukoulas S , et al. , 2016. Monitoring sensitivity to land degradation and Desertification with the environmentally sensitive area index: the case of lesvos island. Land Degradation and Development, 27: 1562-1573. DOI: 10.1002/ldr.2285 .
doi: 10.1002/ldr.2285
Taghizadeh-Mehrjardi R , Minasny B , McBratney AB , et al. , 2012. Digital soil mapping of soil classes using decision trees in Central Iran. In: Minasny B, Malone BP, McBratney AB, (eds.) Digital Soil Assessments and Beyond. Leiden: CRC Press. pp. 197-202.
Torabi HA , Darabi H , Karimidastenaei Z , et al. , 2021. Land degradation risk mapping using topographic, human-induced, and geo-environmental variables and machine learning algorithms, for the Pole-Doab watershed, Iran. Environmental Earth Science, 80. DOI: 10.1007/s12665-020-09327-2 .
doi: 10.1007/s12665-020-09327-2
Trotta C , Menegoni P , Manfredi Frattarelli FM , et al. , 2015. Assessing desertification vulnerability on a local scale: the Castelporziano study case (central Italy). Rendiconti Lincei, 26: 421. DOI:10.1007/s12210-014-0362-5 .
doi: 10.1007/s12210-014-0362-5
UNCCD , 2014. Desertification the Invisible Frontline. .
Vågen TG , Winowiecki LA , Tondoh JE , et al. , 2016. Mapping of soil properties and land degradation risk in Africa using MODIS reflectance. Geoderma, 263: 216-225. DOI: 10. 1016/j.geoderma.2015.06.023 .
doi: 10. 1016/j.geoderma.2015.06.023
Vaysse K , Lagacherie P , 2017. Using quantile regression forest to estimate uncertainty of digital soil mapping products. Geoderma, 291: 55-64. DOI: 10.1016/j.geoderma.2016. 12.017 .
doi: 10.1016/j.geoderma.2016. 12.017
Wang J , Peng J , Li H , et al. , 2021. Soil Salinity Mapping Using Machine Learning Algorithms with the Sentinel-2 MSI in Arid Areas, China. Remote Sensing, 13(2): 305. DOI: 10.3390/rs13020305 .
doi: 10.3390/rs13020305
Wiesmeier M , Barthold F , Blank B , et al. , 2010. Digital mapping of soil organic matter stocks using Random Forest modeling in a semi-arid steppe ecosystem. Plant and Soil, 340: 7-24. DOI: 10.1007/s11104-010-0425-z .
doi: 10.1007/s11104-010-0425-z
Yousefi S , Pourghasemi HR , Avand M , et al. , 2021. Assessment of land degradation using machine-learning techniques: A case of declining rangelands. Land Degradation and Development, 32: 1452-1466. DOI: 10.1002/ldr.3794 .
doi: 10.1002/ldr.3794
No related articles found!
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!