Spatial distribution of aboveground biomass stock in tropical dry forest in Brazil

Abstract

Climate change is being intensified by anthropogenic emission of greenhouse gasses, highlighting the value of forests for carbon dioxide storing carbon in their biomass. Seasonally dry tropical forests are a neglected, threatened, but potentially critical biome for helping mitigate climate change. In South America, knowing the amount and distribution of carbon in Caatinga seasonally dry vegetation is essential to understand its contribution to the global carbon cycle and subsequently design a strategic plan for its conservation. The present study aimed to model and map the spatial distribution of the potential forest biomass stock across 32 forest fragments of Caatinga, in the state of Bahia, Brazil, using regression kriging and Inverse Square of Distance techniques, building from point measurements of vegetation biomass made on-the-ground in ecological plots. First, a model for estimating biomass was fitted as a function of environmental variables to apply regression kriging, and then applied to the maps of the selected components. Elevation, temperature, and precipitation explained 46% of the biomass variations in the Caatinga. The model residuals showed strong spatial dependence and were mapped based on geostatistical criteria, selecting the spherical semivariogram model for interpolation by ordinary kriging. Biomass was also mapped by the Inverse Square of Distance approach. The quality of the regression model suggests that there is good potential for estimating biomass here from environmental variables. The regression kriging showed greater detail in the spatial distribution and revealed a spatial trend of increasing biomass from the north to south of the domain. Additional studies with greater sampling intensity and the use of other explanatory variables are suggested to improve the model, as well as to maximize the technique’s ability to capture the actual biomass behavior in this newly studied seasonally dry ecosystem.

Keywords

• Geostatistics
• Regression Kriging
• Spatial Analysis
• Forest Inventory