Stochastic spatial random forest (SS-RF) for interpolating probabilities of missing land cover data

Jacinta Holloway-Brown,Kate J Helmstedt,Kerrie L Mengersen

doi:10.1186/s40537-020-00331-8

Abstract

Forests are a global environmental priority that need to be monitored frequently and at large scales. Satellite images are a proven useful, free data source for regular global forest monitoring but these images often have missing data in tropical regions due to climate driven persistent cloud cover. Remote sensing and statistical approaches to filling these missing data gaps exist and these can be highly accurate, but any interpolation method results are uncertain and these methods do not provide measures of this uncertainty. We present a new two-step spatial stochastic random forest (SS-RF) method that uses random forest algorithms to construct Beta distributions for interpolating missing data. This method has comparable performance with the traditional remote sensing compositing method, and additionally provides a probability for each interpolated data point. Our results show that the SS-RF method can accurately interpolate missing data and quantify uncertainty and its applicability to the challenge of monitoring forest using free and incomplete satellite imagery data. We propose that there is scope for our SS-RF method to be applied to other big data problems where a measurement of uncertainty is needed in addition to estimates.

Highlights

Forest management is a global priority identified by the United Nations and World Bank Sustainable Development Goals (SDGs)[1] and is important in most countries in the world
In this paper we present a new stochastic spatial random forest (SS-RF) method that addresses both of the gaps identified above and builds on the idea of compositing by using recently observed values of the missing data to produce an interpolated land cover classification
By comparing true land cover classes with the probabilities, we can see that the stochastic random forest (SS-RF) method accurately identifies land cover for this sample (Fig. 7)

Summary

Introduction

Forest management is a global priority identified by the United Nations and World Bank Sustainable Development Goals (SDGs)[1] and is important in most countries in the world. Forest presence can be identified in satellite images for large scale areas by calculating vegetation indices from the spectral bands, which are the colours and near-infrared information captured in the images[2]. Large scale forest cover maps are a transparent tool for identifying forest growth, forest clearing and degradation, and quantifying environmental issues associated with these changes in forest cover such as biodiversity, carbon stocks and social welfare[3]. Examples of these forest cover maps derived from satellite images at global and regional scales include[3,4,5] and[6]. The benefits of using satellite images to construct large scale forest cover maps have been

Methods

Results

Discussion

Conclusion