Digital mapping for cost-effective and accurate prediction of the depth and carbon stocks in Indonesian peatlands

Abstract Tropical peatlands have an important role in the global carbon cycle. In order to quantify carbon stock for peatland management and conservation, the knowledge of the spatial distribution of peat and its depth is essential. This paper proposed a cost-effective and accurate methodology for mapping peat depth and carbon stocks in Indonesia. The method, based on the scorpan spatial soil prediction function framework, was tested in Ogan Komering Ilir, South Sumatra and Katingan, Central Kalimantan. A peat hydrological unit, where a peatland is bounded by at least two rivers, is defined as the mapping area or extent. Peat depth is modelled as a function of topography and spatial position. Four machine learning models were evaluated to model and map peat depth: Cubist regression tree, Random Forests (RF), Quantile Regression Forests (QRF) and Artificial Neural Network (ANN). Covariates representing topography and spatial position were derived from the 1 arc-second digital elevation model (DEM) of the Shuttle Radar Topography Mission (SRTM) (resolution of 30.7 m). The spatial models were calibrated from field observations. For model calibration and uncertainty analysis, the k-fold cross validation approach was used. Three models: Cubist, Random Forests, and Quantile Regression Forests models showed excellent accuracies of peat depth prediction for both areas where the coefficient of determination values range from 0.67 to 0.92 and root mean squared error (RMSE) values range from 0.6 to 1.1 m. ANN showed inferior results. In addition, QRF and Cubist showed the best account of the uncertainty of prediction, in terms of percentage of observations that fall within the defined 90% confidence interval. In terms of the best predictor, elevation comes first. Using the spatial prediction functions, peat depth maps along with their 90% confidence interval were generated. The estimated mean carbon stock for Ogan Komering Ilir is 0.474 Gt and for Katingan is 0.123 Gt. Our estimate for Ogan Komering Ilir is twice larger than a previous study because we mapped the peatland hydrological unit, while the previous study only delineated peat domes. Finally, we recommend a sampling method for peat depth mapping using numerical stratification of elevation to cover both the geographical and covariate space. We expect that the combination of an improved sampling strategy, machine learning models, and kriging will increase the accuracy of peat depth mapping.

[1]  C. Ballabio,et al.  Mapping topsoil physical properties at European scale using the LUCAS database , 2016 .

[2]  Benjamin L Turner,et al.  Environmental controls of temporal and spatial variability in CO2 and CH4 fluxes in a neotropical peatland , 2013, Global change biology.

[3]  Gangcai Liu,et al.  Spatial prediction of soil organic matter content integrating artificial neural network and ordinary kriging in Tibetan Plateau , 2014 .

[4]  Benjamin L Turner,et al.  Contribution of subsurface peat to CO2 and CH4 fluxes in a neotropical peatland , 2011 .

[5]  B. Minasny,et al.  Comparing data mining classifiers to predict spatial distribution of USDA-family soil groups in Baneh region, Iran , 2015 .

[6]  Susan Page,et al.  High diversity of tropical peatland ecosystem types in the Pastaza‐Marañón basin, Peruvian Amazonia , 2011 .

[7]  Budiman Minasny,et al.  Digital soil property mapping and uncertainty estimation using soil class probability rasters , 2015 .

[8]  Alfred E. Hartemink,et al.  Digital mapping of soil carbon in a viticultural region of Southern Brazil , 2016 .

[9]  L. A. Goodman.  On the Exact Variance of Products , 1960 .

[10]  Christopher J. Banks,et al.  Global and regional importance of the tropical peatland carbon pool , 2011 .

[11]  M. Wiesmeier,et al.  Digital mapping of soil organic matter stocks using Random Forest modeling in a semi-arid steppe ecosystem , 2011, Plant and Soil.

[12]  De-Cheng Li,et al.  Mapping soil organic carbon content by geographically weighted regression: A case study in the Heihe River Basin, China , 2016 .

[13]  Limin Peng,et al.  Quantile Regression for Left‐Truncated Semicompeting Risks Data , 2011, Biometrics.

[14]  Susan E. Page,et al.  High diversity of tropical peatland ecosystem types in the Pastaza-Marañón basin, Peruvian Amazonia , 2011 .

[15]  S. Limin,et al.  A cost-efficient method to assess carbon stocks in tropical peat soil , 2012 .

[16]  Budiman Minasny,et al.  A conditioned Latin hypercube method for sampling in the presence of ancillary information , 2006, Comput. Geosci..

[17]  J. McKinley,et al.  Spatial statistics to estimate peat thickness using airborne radiometric data , 2013 .

[18]  Zhi-Wei Liu,et al.  Digital mapping of soil organic matter for rubber plantation at regional scale: An application of random forest plus residuals kriging approach , 2015 .

[19]  B. Huwe,et al.  Uncertainty in the spatial prediction of soil texture: Comparison of regression tree and Random Forest models , 2012 .

[20]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[21]  Jin Zhang,et al.  An overview and comparison of machine-learning techniques for classification purposes in digital soil mapping , 2016 .

[22]  Rudiyanto,et al.  Further results on comparison of methods for quantifying soil carbon in tropical peats , 2016 .

[23]  Susan Page,et al.  A record of Late Pleistocene and Holocene carbon accumulation and climate change from an equatorial peat bog (Kalimantan, Indonesia): implications for past, present and future carbon dynamics , 2004 .

[24]  Bradley A. Miller,et al.  Spatial Modeling of Organic Carbon in Degraded Peatland Soils of Northeast Germany , 2015 .

[25]  D. Charman,et al.  A method for modelling peat depth in blanket peatlands , 2012 .

[26]  Benjamin L Turner,et al.  Tropical wetlands: A missing link in the global carbon cycle? , 2014, Global biogeochemical cycles.

[27]  H. Elsenbeer,et al.  Soil organic carbon concentrations and stocks on Barro Colorado Island — Digital soil mapping using Random Forests analysis , 2008 .

[28]  R. Fyfe,et al.  The importance of sub‐peat carbon storage as shown by data from Dartmoor, UK , 2014 .

[29]  D. Vitt,et al.  Peatland dynamics in a complex landscape: Development of a fen‐bog complex in the Sporadic Discontinuous Permafrost zone of northern Alberta, Canada , 2011 .

[30]  Mathias Disney,et al.  Carbon storage in peatlands: A case study on the Isle of Man , 2013 .

[31]  Pete Smith,et al.  Assessing existing peatland models for their applicability for modelling greenhouse gas emissions from tropical peat soils , 2011 .

[32]  C. Akumu,et al.  Modeling peatland carbon stock in a delineated portion of the Nayshkootayaow river watershed in Far North, Ontario using an integrated GIS and remote sensing approach , 2014 .

[33]  M. Larocque,et al.  Determining the number of manual measurements required to improve peat thickness estimations by ground penetrating radar , 2009 .

[34]  André St-Hilaire,et al.  A review of ground-penetrating radar studies related to peatland stratigraphy with a case study on the determination of peat thickness in a northern boreal fen in Quebec, Canada , 2013 .

[35]  A. McBratney,et al.  A continuum approach to soil classification by modified fuzzy k‐means with extragrades , 1992 .

[36]  B. D. Ripley,et al.  Neural Networks for Pattern Recognition.@@@Pattern Recognition and Neural Networks. , 1997 .

[37]  A. McBratney,et al.  Further results on prediction of soil properties from terrain attributes: heterotopic cokriging and regression-kriging , 1995 .

[38]  Budiman Minasny,et al.  On digital soil mapping , 2003 .

[39]  Rudiyanto,et al.  Estimating distribution of carbon stock in tropical peatland using a combination of an Empirical Peat Depth Model and GIS , 2015 .

[40]  Andy Liaw,et al.  Classification and Regression by randomForest , 2007 .

[41]  Florian Siegert,et al.  Determination of the amount of carbon stored in Indonesian peatlands. , 2008 .

[42]  J. Esterle,et al.  Spatial variability in modern tropical peat deposits from Sarawak, Malaysia and Sumatra, Indonesia: analogues for coal , 1994 .

[43]  Malcolm Coull,et al.  Mapping soil carbon stocks across Scotland using a neural network model , 2016 .

[44]  J. Holden,et al.  Evaluating approaches for estimating peat depth , 2014 .

[45]  Pete Smith,et al.  Comparison of methods for quantifying soil carbon in tropical peats , 2014 .

[46]  Margaret G. Schmidt,et al.  Predictive soil parent material mapping at a regional-scale: a Random Forest approach. , 2014 .

[47]  Ian D. Moore,et al.  Terrain attributes: estimation methods and scale effects , 1993 .

[48]  Nicolai Meinshausen,et al.  Quantile Regression Forests , 2006, J. Mach. Learn. Res..

[49]  M. A. Brady.  Organic matter dynamics of coastal peat deposits in Sumatra, Indonesia , 1997 .

[50]  J. Ross Quinlan,et al.  Combining Instance-Based and Model-Based Learning , 1993, ICML.

[51]  W. Peizhuang.  Pattern Recognition with Fuzzy Objective Function Algorithms (James C. Bezdek) , 1983 .

[52]  Christopher M. Bishop,et al.  Neural networks for pattern recognition , 1995 .

[53]  Budiman Minasny,et al.  High‐Resolution 3‐D Mapping of Soil Texture in Denmark , 2013 .

[54]  H. Vereecken,et al.  Mapping peat layer properties with multi-coil offset electromagnetic induction and laser scanning elevation data , 2016 .

[55]  A. Avagyan,et al.  Towards climate-responsible peatlands management. , 2014 .

[56]  Hecht-Nielsen.  Theory of the backpropagation neural network , 1989 .

[57]  D. Vitt,et al.  Development of a peatland complex in boreal western Canada: lateral site expansion and local variability in vegetation succession and long-term peat accumulation , 2003 .

[58]  J. A. Hartigan,et al.  A k-means clustering algorithm , 1979 .

[59]  Y. Bergeron,et al.  Quantifying spatial and temporal Holocene carbon accumulation in ombrotrophic peatlands of the Eastmain region, Quebec, Canada , 2011 .

[60]  D. Charman,et al.  Modelling soil organic carbon distribution in blanket peatlands at a landscape scale , 2013 .

[61]  Helaina I. J. Black,et al.  Predicting Scottish topsoil organic matter content from colour and environmental factors , 2015 .

[62]  G. Heuvelink,et al.  A generic framework for spatial prediction of soil variables based on regression-kriging , 2004 .

[63]  A. Bauriegel,et al.  Estimating vulnerability of agriculturally used peatlands in north-east Germany to carbon loss based on multi-temporal subsidence data analysis , 2016 .

[64]  Alex B. McBratney,et al.  Spatial prediction of soil properties from landform attributes derived from a digital elevation model , 1994 .

[65]  Stefan Fritsch,et al.  neuralnet: Training of Neural Networks , 2010, R J..

[66]  W Shotyk,et al.  Interdependence of peat and vegetation in a tropical peat swamp forest. , 1999, Philosophical transactions of the Royal Society of London. Series B, Biological sciences.

[67]  Patrick Bogaert,et al.  Updating soil survey maps using random forest and conditioned Latin hypercube sampling in the loess derived soils of northern Iran , 2014 .

[68]  M. Kaneko,et al.  The carbon content characteristics of tropical peats in Central Kalimantan, Indonesia: Estimating their spatial variability in density , 2001 .

[69]  J. Mohammadi,et al.  Spatial Estimation of Saturated Hydraulic Conductivity from Terrain Attributes Using Regression, Kriging, and Artificial Neural Networks , 2011 .

[70]  J. Ross Quinlan,et al.  C4.5: Programs for Machine Learning , 1992 .

[71]  Budiman Minasny,et al.  Mapping continuous depth functions of soil carbon storage and available water capacity , 2009 .

[72]  James C. Bezdek,et al.  Pattern Recognition with Fuzzy Objective Function Algorithms , 1981, Advanced Applications in Pattern Recognition.

[73]  Geoffrey H. Ball,et al.  ISODATA, A NOVEL METHOD OF DATA ANALYSIS AND PATTERN CLASSIFICATION , 1965 .

[74]  J. Anderson.  Ecology and forest types of the peat swamp forests of Sarawak and Brunei in relation to their silviculture , 1961 .

[75]  N. Holden,et al.  Estimating the carbon stock of a blanket peat region using a peat depth inference model , 2011 .

[76]  Manuel A. Aguilar,et al.  Effects of Terrain Morphology, Sampling Density, and Interpolation Methods on Grid DEM Accuracy , 2005 .

[77]  José Augusto Baranauskas,et al.  How Many Trees in a Random Forest? , 2012, MLDM.

[78]  Budiman Minasny,et al.  Digital soil mapping: A brief history and some lessons , 2016 .