Random Forest Variable Importance Spectral Indices Scheme for Burnt Forest Recovery Monitoring - Multilevel RF-VIMP

Burnt forest recovery is normally monitored with a time-series analysis of satellite data because of its proficiency for large observation areas. Traditional methods, such as linear correlation plotting, have been proven to be effective, as forest recovery naturally increases with time. However, these methods are complicated and time consuming when increasing the number of observed parameters. In this work, we present a random forest variable importance (RF-VIMP) scheme called multilevel RF-VIMP to compare and assess the relationship between 36 spectral indices (parameters) of burnt boreal forest recovery in the Great Xing’an Mountain, China. Six Landsat images were acquired in the same month 0, 1, 4, 14, 16, and 20 years after a fire, and 39,380 fixed-location samples were then extracted to calculate the effectiveness of the 36 parameters. Consequently, the proposed method was applied to find correlations between the forest recovery indices. The experiment showed that the proposed method is suitable for explaining the efficacy of those spectral indices in terms of discrimination and trend analysis, and for showing the satellite data and forest succession dynamics when applied in a time series. The results suggest that the tasseled cap transformation wetness, brightness, and the shortwave infrared bands (both 1 and 2) perform better than other indices for both classification and monitoring.

[1]  E. Chuvieco,et al.  Burn severity estimation from remotely sensed data: Performance of simulation versus empirical models , 2007 .

[2]  Roberta E. Martin,et al.  A Tale of Two “Forests”: Random Forest Machine Learning Aids Tropical Forest Carbon Mapping , 2014, PloS one.

[3]  A. Kozak,et al.  Effects of multicollinearity and autocorrelation on the variable-exponent taper functions , 1997 .

[4]  Wendy Francesconi,et al.  Environmental predictors of forest change: An analysis of natural predisposition to deforestation in the tropical Andes region, Peru , 2018 .

[5]  J. Pereira,et al.  Modeling spatial patterns of fire occurrence in Mediterranean Europe using Multiple Regression and Random Forest , 2012 .

[6]  A. Gitelson,et al.  Use of a green channel in remote sensing of global vegetation from EOS- MODIS , 1996 .

[7]  Chengquan Huang,et al.  USE OF MULTIPLE SPECTRAL INDICES TO ESTIMATE BURN SEVERITY IN THE BLACK HILLS OF SOUTH DAKOTA , 2008 .

[8]  R. H. Myers.  Classical and modern regression with applications , 1986 .

[9]  Achim Zeileis,et al.  Bias in random forest variable importance measures: Illustrations, sources and a solution , 2007, BMC Bioinformatics.

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

[11]  Cecilia Masemola.  Remote sensing of leaf area index in Savannah grass using inversion of radiative transfer model on Landsat 8 imagery : case study Mpumalanga, South Africa , 2015 .

[12]  D. Kleinbaum,et al.  Applied Regression Analysis and Other Multivariate Methods , 1978 .

[13]  Achim Zeileis,et al.  A New, Conditional Variable-Importance Measure for Random Forests Available in the party Package , 2009 .

[14]  Robert P. Sheridan,et al.  Random Forest: A Classification and Regression Tool for Compound Classification and QSAR Modeling , 2003, J. Chem. Inf. Comput. Sci..

[15]  D. Griffith.  Spatial Autocorrelation: A Primer , 1987 .

[16]  Limin Yang,et al.  Derivation of a tasselled cap transformation based on Landsat 7 at-satellite reflectance , 2002 .

[17]  Tetsuro Sakai,et al.  Extraction of burned forest area in the Greater Hinggan Mountain of China based on Landsat TM data , 2013, 2013 IEEE International Geoscience and Remote Sensing Symposium - IGARSS.

[18]  S. Ollinger.  Sources of variability in canopy reflectance and the convergent properties of plants. , 2011, The New phytologist.

[19]  R. Sokal,et al.  Applications of Spatial Autocorrelation in Ecology , 1987 .

[20]  Trevor Hastie,et al.  An Introduction to Statistical Learning , 2013, Springer Texts in Statistics.

[21]  B. Gao.  NDWI—A normalized difference water index for remote sensing of vegetation liquid water from space , 1996 .

[22]  R. Sokal,et al.  Spatial autocorrelation in biology: 1. Methodology , 1978 .

[23]  A. Huete,et al.  Overview of the radiometric and biophysical performance of the MODIS vegetation indices , 2002 .

[24]  Zhihua Liu.  Effects of climate and fire on short-term vegetation recovery in the boreal larch forests of Northeastern China , 2016, Scientific Reports.

[25]  K. Hornik,et al.  party : A Laboratory for Recursive Partytioning , 2009 .

[26]  S. K. McFeeters.  The use of the Normalized Difference Water Index (NDWI) in the delineation of open water features , 1996 .

[27]  V. Shandas,et al.  Assessing the Potential of Land Use Modification to Mitigate Ambient NO2 and Its Consequences for Respiratory Health , 2017, International journal of environmental research and public health.

[28]  T. Mitchell Aide,et al.  Consequences of the Armed Conflict, Forced Human Displacement, and Land Abandonment on Forest Cover Change in Colombia: A Multi-scaled Analysis , 2013, Ecosystems.

[29]  A. Huete,et al.  Development of a two-band enhanced vegetation index without a blue band , 2008 .

[30]  P. Moran.  Notes on continuous stochastic phenomena. , 1950, Biometrika.

[31]  P. Fulé,et al.  Comparison of burn severity assessments using Differenced Normalized Burn Ratio and ground data , 2005 .

[32]  G. Rondeaux,et al.  Optimization of soil-adjusted vegetation indices , 1996 .

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

[34]  Koreen Millard,et al.  On the Importance of Training Data Sample Selection in Random Forest Image Classification: A Case Study in Peatland Ecosystem Mapping , 2015, Remote. Sens..

[35]  Achim Zeileis,et al.  BMC Bioinformatics BioMed Central Methodology article Conditional variable importance for random forests , 2008 .

[36]  Yongqiang Liu,et al.  Chapter 13 Remote Sensing Applications of Wildland Fire and Air Quality in China , 2008 .

[37]  W. Cohen,et al.  Testing a Landsat-based approach for mapping disturbance causality in U.S. forests , 2017 .

[38]  Xulin Guo,et al.  Remote Sensing Techniques in Monitoring Post-Fire Effects and Patterns of Forest Recovery in Boreal Forest Regions: A Review , 2013, Remote. Sens..

[39]  Feng Zhao,et al.  Using High Spatial Resolution Satellite Imagery to Map Forest Burn Severity Across Spatial Scales in a Pine Barrens Ecosystem , 2017 .

[40]  Anping Chen,et al.  Early post-fire regeneration of a fire-prone subtropical mixed Yunnan pine forest in Southwest China: Effects of pre-fire vegetation, fire severity and topographic factors , 2015 .

[41]  C. Beijing.  Changes of Forest Fire Danger and the Evaluation of the FWI System Application in the Daxing’anling Region , 2010 .

[42]  J. G. White,et al.  Aerial Color Infrared Photography for Determining Early In‐Season Nitrogen Requirements in Corn , 2005 .

[43]  A. Gitelson,et al.  Novel algorithms for remote estimation of vegetation fraction , 2002 .

[44]  V. Caselles,et al.  Mapping burns and natural reforestation using thematic Mapper data , 1991 .

[45]  Guosheng Zhong,et al.  Long-Term Satellite Detection of Post-Fire Vegetation Trends in Boreal Forests of China , 2013, Remote. Sens..

[46]  J. Evans,et al.  Modeling Species Distribution and Change Using Random Forest , 2011 .

[47]  Nikos Koutsias,et al.  A Remote Sensing and GIS Approach to Study the Long-Term Vegetation Recovery of a Fire-Affected Pine Forest in Southern Greece , 2015, Remote. Sens..