NARR data analysis identifies climate change anomalies in Ontario's managed boreal forest

Keywords: climate change, spatial analysis, geographic information systems, multi-level data
Abstract Type: Paper Abstract

Abstract

The United States National Oceanic and Atmospheric Administration’s North American Regional Reanalysis (NARR) dataset has proven useful in studies assessing climate and climate-related variables. We implemented forty years of 0.3° × 0.3° NARR data in spatial analysis to identify canopy conductance anomalies, significant differences between conductance observed during the defined climate normal (1980-2010) and change (2011-2020) periods, within Ontario's 30 Mha of managed boreal forest. Since the mid-1900s, humans have actively accelerated Earth’s global warming by increasing anthropogenic greenhouse gas emissions, resulting in variations in climate systems operating over vast extents. Canada's mean temperature is rising at a rate about double that of the global decadal average of 0.1-0.2°C. Systematic climate examinations over vast extents and long periods provide a litmus test for statistically significant variations. However, datasets of environmental variables covering large areas are known to exhibit significant spatial autocorrelation consequent to directional anisotropy emphasized by geostatistical interpolation methods and properties of remotely sensed reflectance. 1st-order polynomial directional trends were examined by Generalized Linear Regression (GLR) to assess the influence of spatial autocorrelation in NARR data for canopy conductance and four conductance parameters. Raw data and 1st-order trends were modeled fitted for GLR with Multiplicative Inverse and Box-Cox transformations and detrended for analysis of variance. Significant statistical differences were identified between the climate normal and change periods for canopy conductance using the two-sample Kolmogorov-Smirnov test. Principal component analysis (PCA) performed on canopy conductance parameters showed that conductance anomalies have clear spatial correlations with climate and climate-related ecological site characteristics.