Changes in Landscape Greenness and Climatic Factors over 25 Years (1989-2013) in the USA.

Maliha Nash,Timothy Wade,Jay Christensen,James Wickham

doi:10.3390/rs9030295

Abstract

Monitoring and quantifying changes in vegetation cover over large areas using remote sensing can be achieved using the Normalized Difference Vegetation Index (NDVI), an indicator of greenness. However, distinguishing gradual shifts in NDVI (e.g., climate related-changes) versus direct and rapid changes (e.g., fire, land development) is challenging as changes can be confounded by time-dependent patterns, and variation associated with climatic factors. In the present study, we leveraged a method that we previously developed for a pilot study to address these confounding factors by evaluating NDVI change using autoregression techniques that compare results from univariate (NDVI vs. time) and multivariate analyses (NDVI vs. time and climatic factors) for 7,660,636 1 km × 1 km pixels comprising the 48 contiguous states of the USA, over a 25-year period (1989–2013). NDVI changed significantly for 48% of the nation over the 25-year period in the univariate analyses where most significant trends (85%) indicated an increase in greenness over time. By including climatic factors in the multivariate analyses of NDVI over time, the detection of significant NDVI trends increased to 53% (an increase of 5%). Comparisons of univariate and multivariate analyses for each pixel showed that less than 4% of the pixels had a significant NDVI trend attributable to gradual climatic changes while the remainder of pixels with a significant NDVI trend indicated that changes were due to direct factors. While most NDVI changes were attributable to direct factors like wildfires, drought or flooding of agriculture, and tree mortality associated with insect infestation, these conditions may be indirectly influenced by changes in climatic factors.

Highlights

Remote sensing data have been used by numerous researchers to monitor and quantify changes in vegetation cover over large areas for long-term time frames [1,2,3,4,5,6,7,8]
While most Normalized Difference Vegetation Index (NDVI) changes were attributable to direct factors like wildfires, drought or flooding of agriculture, and tree mortality associated with insect infestation, these conditions may be indirectly influenced by changes in climatic factors
Distinguishing gradual shifts in NDVI versus direct and rapid changes is challenging as changes can be confounded by variation associated with climatic factors and by time-dependent patterns

Summary

Introduction

Remote sensing data have been used by numerous researchers to monitor and quantify changes in vegetation cover over large areas for long-term time frames [1,2,3,4,5,6,7,8]. NDVI has been used to identify gradual changes over decades [18,19], as well as Remote Sens. Distinguishing gradual shifts in NDVI (e.g., climate-related changes) versus direct and rapid changes (e.g., fire, land development) is challenging as changes can be confounded by variation associated with climatic factors and by time-dependent patterns. Climatic factors such as precipitation and temperature often strongly influence vegetation physiology and phenology and greenness [23,24]. Climatic factors may show a general pattern of change over time [30], and may account for a trend in NDVI in some areas. To detect change in vegetation cover, it is important to account for time-dependent patterns in NDVI, which are typically pronounced [5,29,31]

Objectives

Methods

Results

Conclusion