Introduction

AGB

Above-ground biomass (AGB) is the total mass of material stored in the living stems, branches and leaves of vegetation, and is often described as a biomass density, with units of mass per unit area such as t/ha.

There have been major efforts to map forest AGB using Earth Observation (EO) data, and further maps are anticipated in the near-future because of new missions dedicated to measuring AGB.

Current AGB maps were derived using different methods and data sources, leading to significant disagreements between them that reduce their value in specific applications such as in estimating global and national carbon stocks. In addition, the maps have specific individual error properties, rendering them unreliable for biomass change analysis, despite representing different epochs.

Plot Data

Assessment of the accuracy of an AGB map has to take into account random and bias errors in the map itself and in the reference data used to validate it. Reference data are commonly used in situ plot measurements (plot data), whose uncertainty can be quantified using specific AGB validation protocols. While plot accuracy and uncertainty at local and regional scales have been addressed, at continental and global scale AGB maps are hampered by the lack of a global reference dataset.

Plot Uncertainty

The consequent lack of a consistently sampled reference AGB dataset has consequences for statistical inference, which is only possible under certain assumptions and requires the data to be accompanied by uncertainty estimates. Using a collection of plot datasets with uncertainty estimates across the globe offers several opportunities such as: revealing regional patterns that may be explainable by environmental and/or ecological variables; investigate bias and develop bias reduction methods; and finally in evaluating of the extent to which plot-map differences can be attributed to map error.

Map Bias

A key principle for climate monitoring is that random errors and time-dependent biases in satellite EO and derived products should be identified. More generally map users prefer AGB maps to be unbiased and to have spatially explicit uncertainty information, including its spatial autocorrelation.

The Plot2Map Package

To address these issues, a comprehensive framework for assessing and harmonizing global AGB maps has been developed with this package. This framework, incorporated with this package, aims to quantify and reduce biases in AGB maps while providing spatially explicit uncertainty information.

The Plot2Map framework incorporates three key components:

1. Plot data preparation: A global collection of plot data is carefully selected and pre-processed to minimize temporal and spatial mismatches with AGB maps.

2. Uncertainty quantification: The framework assesses various sources of uncertainty in plot-level AGB estimates, including measurement errors, allometric model errors, and sampling errors.

3. Map bias assessment and correction: Using the prepared plot data, the framework quantifies biases in global AGB maps and models these biases using spatial covariates.

Installation

You can install Plot2Map from GitHub:

# install.packages("devtools")
devtools::install_github("aTnT/Plot2Map")

library(Plot2Map)

Vignettes

In the following vignettes, each of the Plot2Map framework components will be addressed by showcasing practical examples of a typical workflow:

1. Plot data preparation.
2. Uncertainty quantification.
3. Map bias assessment and correction.