Publications

Gatis, N; Anderson, K; Grand-Clement, E; Luscombe, DJ; Hartley, IP; Smith, D; Brazier, RE (2017). Evaluating MODIS vegetation products using digital images for quantifying local peatland CO2 gas fluxes. REMOTE SENSING IN ECOLOGY AND CONSERVATION, 3(4), 217-231.

Abstract
In peatlands plant growth and senescence affect annual ecosystem carbon dioxide (CO2) exchange, and CO2 fluxes show considerable inter-annual variability. Phenology is a fundamental indicator of ecosystem carbon dynamics and can be measured from remote sensing systems, but the extent to which satellite products provide useful proxies of peatland vegetation phenology is not well known. Using MODIS vegetation products coupled with field observations of phenology from a basic camera system and measurements of Gross Primary Productivity (GPP) measured using a closed chamber method, we sought to establish the extent to which satellite observations capture phenological processes at a UK peatland site. Daily, true-colour digital images were captured with a time-lapse camera (Brinno TLC100) between 23-Apr-2013 and 29-Oct-2013 and converted into a Green-Red Vegetation Index (GRVI). These were compared with a range of MODIS vegetation products at various spatial resolutions. We found that vegetation products with finer spatial resolution (<500m) more accurately captured spring green-up (e.g. Normalized Difference Vegetation Index 16-day product), whereas those with 8-day temporal resolution better captured whole-season dynamics. The 8-day Gross Primary Productivity (GPP8) and the fraction of absorbed photosynthetically active radiation (fPAR8) products had the strongest daily Pearson's correlations with camera-derived GRVI (r>0.90). The camera-GRVI (P=0.005, r=0.98) and MODIS-GRVI (P=0.041, r=0.89) products showed the strongest significant correlations with gross primary productivity measured using static chambers in the field. This work demonstrates that freely available MODIS data captured up to 92% of the daily variation in phenology over an upland peatland. This approach shows great potential for capturing ecosystem carbon dynamics which underpin carbon trading schemes, a budding funding source for peatland restoration projects.

DOI:
10.1002/rse2.45

ISSN:
2056-3485