Prediction of Moisture, Carbon, and Nitrogen in Large Pristine Peatlands of Central Canada by NIRS

[Introduction] Peatlands around the globe are receiving increasing attention for impacts of climate change, such as drying and the increasing risk of intractable fires. Peatlands represent a major store of the world’s land-based carbon. They cover an estimated 3 - 4% of the earth’s land surface and store 25% of the world’s terrestrial carbon, roughly equivalent to 75% of the total amount of global atmospheric C. Thus, the fate of peat-lands on the globe can influence concentrations of greenhouse gases in the atmosphere. Peatlands are most concentrated in the northern hemisphere and are an asset and a responsibility largely of northern nations. North America has about 40% of the world's peatlands with a total area of 1,735,000 km². Peatlands occur in every province of Canada and in nearly every state of the United States. They cover 12.3%, i.e., 1,113,270 km2 in Canada and contain over 150 Giga tonnes of carbon, about 60% of Canada’s carbon stock. Yet monitoring is fragmented geographically and in time. Near-infrared spectroscopy offers an on-site available, and economical means for monitoring peatlands.

[Materials and Methods]
Five sampling sites were located in diverse areas of the 810,000 ha Traditional Territory of the Poplar River First Nation on the east side of Lake Winnipeg, Manitoba. Peat was cored using the Russian Peat Borer, a 50-cm long side-filling core sampler, with extension rods to permit the collection of samples at deeper levels. Individual sub-samples were 10 cm in length. Spectra were collected from field-moist samples using the Zeiss Corona® 45 VISNIR from 380 to 1690 nm. Samples were analyzed for moisture, carbon, and nitrogen. Moisture was determined by drying samples at 90º C for 8 to 16 hours, depending on the moisture content, until they reached constant weight. Analyses for carbon and nitrogen were performed by the Ontario Forest Research Institute using a CHN Analyzer. Calibration models between the spectral and reference data were developed in The Unscrambler using principal component analysis/partial least squares regression (PCA/PLS).

[Results and Discussion]
Calibrations developed for moisture had r2 ranging from 0.82 to 0.97. For carbon, r2 averaged 0.87 and for nitrogen, 0.77. In conclusion, NIR technology is effective for the rapid determination of moisture and carbon, in field-moist samples from peatlands. The technology is a potential tool for enhancing knowledge of carbon inventories and carbon quality in peatlands and for spatial and temporal monitoring of responses of peatlands to climate change. Using field-portable instruments, peat analysis can be conducted in remote areas, reducing costs of sample handling and storage.