Jun 8, 2012
'Better' data show no change in Amazon greenness for 2000–2009
Scientists using satellite data to investigate greenness dynamics should take more care to filter out all corrupt data before drawing any conclusions, according to researchers in the US.
In a paper in Environmental Research Letters, the group outlines a number of analyses that show how corrupt satellite data for Amazon forests have given anomalous results in the past.
Lead author Arindam Samanta from Boston University, now at Atmospheric and Environmental Research (AER), US, told environmentalresearchweb: "Researchers should take the time to filter out all the data that has been corrupted by atmospheric effects such as clouds and aerosols (e.g. from burning biomass). Many seem to think that the satellite sensor automatically adjusts for these effects, but this is not true."
In their study, which focuses on Amazon forests in the 2000–2009 time period, Samanta and his colleagues estimated biases resulting from screening corrupted data from analysis; analysed characteristics of greenness dynamics, taking into consideration the statistical distributions of measurement errors of the greenness data; investigated the relationship between greenness and precipitation anomalies; and investigated the persistence of greenness changes over time.
When looking at data from NASA's Moderate Resolution Imaging Spectroradiometer (MODIS) for the Amazon forests, Samanta and his colleagues found that roughly 60–66% of the area affected by drought in 2005 lacks uncorrupted enhanced vegetation index (EVI) data.
"Given the selective prevalence of clouds and aerosols, the uncorrupted data are not representative of the larger drought area, and therefore greenness dynamics inferred from the available sample cannot be extrapolated to the larger area," said Samanta.
A decade's worth of Terra MODIS Collection 5 EVI data and six years of a previous version, both analysed in multiple ways and taking into account EVI accuracy, consistently show a pattern of negligible changes in the greenness levels of forests both in the area affected by drought in 2005 and outside it. The researchers conclude that there were no changes in the greenness of these forests, or if there were changes, the EVI data failed to capture them.
"These findings contradict findings by other research groups but we believe our data is more reliable because we have applied all the necessary filters and quality flags," said Samanta. "This analysis demonstrates that there is a need to evaluate rigorously satellite-measured greenness data before utilizing in interpretation of vegetation greenness changes. In particular, atmospheric influences should be properly screened out, and any resulting biases should be well understood."
Samanta and his colleagues hope that their data-analysis techniques will be useful to researchers around the world and enable them to analyse satellite data more accurately.
About the author
Nadya Anscombe is a freelance science journalist based in Bristol, UK.