A new NDVI measure that overcomes data sparsity in cloud-covered regions predicts annual variation in ground-based estimates of high arctic plant productivity

Stein Rune Karlsen,Helen B Anderson,René Van Der Wal,Brage Bremset Hansen

doi:10.1088/1748-9326/aa9f75

Abstract

Efforts to estimate plant productivity using satellite data can be frustrated by the presence of cloud cover. We developed a new method to overcome this problem, focussing on the high-arctic archipelago of Svalbard where extensive cloud cover during the growing season can prevent plant productivity from being estimated over large areas. We used a field-based time-series (2000−2009) of live aboveground vascular plant biomass data and a recently processed cloud-free MODIS-Normalised Difference Vegetation Index (NDVI) data set (2000−2014) to estimate, on a pixel-by-pixel basis, the onset of plant growth. We then summed NDVI values from onset of spring to the average time of peak NDVI to give an estimate of annual plant productivity. This remotely sensed productivity measure was then compared, at two different spatial scales, with the peak plant biomass field data. At both the local scale, surrounding the field data site, and the larger regional scale, our NDVI measure was found to predict plant biomass (adjusted R2 = 0.51 and 0.44, respectively). The commonly used ‘maximum NDVI’ plant productivity index showed no relationship with plant biomass, likely due to some years having very few cloud-free images available during the peak plant growing season. Thus, we propose this new summed NDVI from onset of spring to time of peak NDVI as a proxy of large-scale plant productivity for regions such as the Arctic where climatic conditions restrict the availability of cloud-free images.

Highlights

The use of satellite-derived data, such as the Normalised Difference Vegetation Index (NDVI), has provided great opportunities for large-scale monitoring of environmental change, in regions that are remote and difficult to access (Kerr and Ostrovsky 2003, Turner et al 2003, Pettorelli et al 2005)
Since several years in our study period were characterized by frequent cloud cover before and during the peak growing season, we developed a new NDVI measure, which calculated the integrated NDVI from the onset (O) to the Peak (P) of the growing season
OP NDVI showed a strong positive relationship with the plant biomass data measured in the field (figure 2(a), figure 2(b); Pearson’s correlation coefficients r = 0.75 and 0.71 for the local and regional mask, respectively), and the relationship was even stronger for the annual rates of change

Summary

Introduction

The use of satellite-derived data, such as the Normalised Difference Vegetation Index (NDVI), has provided great opportunities for large-scale monitoring of environmental change, in regions that are remote and difficult to access (Kerr and Ostrovsky 2003, Turner et al 2003, Pettorelli et al 2005). There can be major drawbacks in deriving accurate assessments of plant productivity over large areas using proxies such as NDVI: for example, the lack of sufficient cloud-free data and the absence of validation with field-based plant biomass data. Very few large-scale studies have validated satellite-derived plant productivity proxies with plant biomass data from field-based studies, with those that have using data from only one season and failing to capture annual variability (Epstein et al 2012, Raynolds et al 2012). To provide more accurate proxies of plant productivity at large spatial scales in regions where cloud cover is prevalent, new methodologies are required which are validated against data from long-term field-based studies

Methods

Results

Discussion

Conclusion