Canadian Forest Service Publications

Choice of satellite imagery and attribution of changes to disturbance type strongly affects forest carbon balance estimates. 2015. Mascorro, V.S.; Coops, N.C.; Kurz, W.A.; Olguin, M. Carbon Balance Manage 10:30

Year: 2015

Available from: Pacific Forestry Centre

Catalog ID: 36523

Language: English

CFS Availability: PDF (download)

Available from the Journal's Web site.
DOI: 10.1186/s13021-015-0041-6

† This site may require a fee.

Abstract

Background: Remote sensing products can provide regular and consistent observations of the Earth´s surface to monitor and understand the condition and change of forest ecosystems and to inform estimates of terrestrial carbon dynamics. Yet, challenges remain to select the appropriate satellite data source for ecosystem carbon monitoring. In this study we examine the impacts of three attributes of four remote sensing products derived from Landsat, Landsat- SPOT, and MODIS satellite imagery on estimates of greenhouse gas emissions and removals: (1) the spatial resolution (30 vs. 250 m), (2) the temporal resolution (annual vs. multi-year observations), and (3) the attribution of forest cover changes to disturbance types using supplementary data. Results: With a spatially-explicit version of the Carbon Budget Model of the Canadian Forest Sector (CBM-CFS3), we produced annual estimates of carbon fluxes from 2002 to 2010 over a 3.2 million ha forested region in the Yucatan Peninsula, Mexico. The cumulative carbon balance for the 9-year period differed by 30.7 million MgC (112.5 million Mg CO2e) among the four remote sensing products used. The cumulative difference between scenarios with and without attribution of disturbance types was over 5 million Mg C for a single Landsat scene. Conclusions: Uncertainty arising from activity data (rates of land-cover changes) can be reduced by, in order of priority, increasing spatial resolution from 250 to 30 m, obtaining annual observations of forest disturbances, and by attributing land-cover changes by disturbance type. Even missing a single year in the land-cover observations can lead to substantial errors in ecosystems with rapid forest regrowth, such as the Yucatan Peninsula.

Plain Language Summary

Quantifying the emissions and removals of greenhouse gases in forests requires data on the rates of disturbances and human impacts. These activity data can be derived from time series of remote-sensing based land cover and land-cover change products. This study examined the impacts of three attributes of four remote sensing products derived from Landsat, Landsat-SPOT, and MODIS satellite imagery on estimates of greenhouse gas emissions and removals: (1) the spatial resolution (30 m vs. 250 m), (2) the temporal resolution (annual vs. multi-year observations), and (3) the attribution of forest cover changes to disturbance types using supplementary data. Analyses with the Carbon Budget Model of the Canadian Forest Sector (CBM-CFS3) estimated that the cumulative carbon balance of a 3.2 million hectare forested region in the Yucatan Peninsula, Mexico for the 9-year period (2002 – 2010) differed by 30.7 million MgC (112.5 million Mg CO2e) among the four remote sensing products used. The cumulative difference between scenarios with and without attribution of disturbance types was over 5 million MgC. Uncertainty arising from activity data can be reduced by, increasing spatial resolution from 250 to 30 meters, obtaining annual observations of forest disturbances, and by attributing land-cover changes by disturbance type. Even missing a single year in the land-cover observations can lead to substantial errors in ecosystems with rapid forest regrowth, such as the Yucatan.

Date modified: