In this paper we describe remote sensing methods with optical and radar sources to monitor aboveground carbon in forests. We move from forest/non-forest mapping, through forest type mapping (conifer, deciduous and mixed) to major species mapping. Biomass combined with species is required to compute the aboveground carbon. Biomass for forests needs accurate height measurements and allometric equations relating height, species and biomass. LIDAR data and TanDEM-X data were used for height measurement. We close with a discussion of accuracies and sources of error in making carbon maps with remote sensing and allometric models.

With increasing CO2 in the atmosphere due to fossil fuel burning, there is a need to quantitatively measure the aboveground carbon in forests. The best remote sensing sensors for this task in Canada are hyperspectral sensors to obtain major forest species, and lidar to measure tree height (H). The Greater Victoria Watershed District on Vancouver Island was selected as a test site and imaged with airborne AVIRIS 4m data and AISA 2m data. Fifty-four ground plots provided excellent ground reference data. Knowing the species, tree heights, and allometric equations relating to these species permits us to determine the aboveground carbon. This paper discusses these measurements, and the variation in carbon estimates due to errors of tree height and species classification.

The world's forests generate oxygen and store carbon, mitigating global climate change. Monitoring the health of these forests is an international priority, which benefits from the use of remote sensing. With hyperspectral sensors capable of discerning forest species and foliar chemistry, many forest information products can be generated including maps of forest species, canopy chemistry, biomass, and carbon. Methods and challenges related to forest monitoring and mapping are discussed, with reference to our past and current work in the field of hyperspectral imaging, and with an emphasis on biomass and carbon mapping. The processes explained in this paper have been tested primarily on a study site on Vancouver Island, Canada.

Hyperspectral sensing of forest chemistry can provide indicators of forest health. Foliar pigments are directly involved with the photo synthetic process and, therefore, are intimately tied to vegetation vigor [1]. Data from the University of Victoria's Airborne Imaging Spectrometer for Applications (AISA) were acquired in 2006 for the Greater Victoria Watershed District (GVWD). Minimum Noise Fraction (MNF) [2], an algorithm based on linear principal components, and a nonlinear local geometric projection algorithm (NL-LGP) [3] were used to denoise this hyperspectral dataset. The initial reflectance and the two denoised datasets were used to generate estimates of foliar chemistry, which were evaluated with field measurements taken prior to the AISA acquisition. Initial analysis was performed at the plot level. Data that were denoised produced marginally less accurate chlorophyll-a estimates than the original data set. The NL-LGP denoising algorithm provided better chemistry mapping at finer spatial resolutions than MNF or the original dataset.

Hyperspectral sensing of forest chemistry can provide indicators of forest health. Foliar pigments are directly involved with the photosynthetic process and, therefore, are intimately tied to vegetation vigor. AISA and AVIRIS hyperspectral datasets were acquired over the Greater Victoria Watershed District test site in 2006 and 2002, respectively. AISA was calibrated to AVIRIS to facilitate sensor comparison. The data were used to generate a forest species classification, endmember fractions and chemistry for test plots. The hyperspectral products were used to separate ground cover (Salal) from the forest overstory and chemistry was estimated for both layers. Classification accuracies exceeded 89% in mapping major forest species. AVIRIS predicted chemistry agreed with measured chemistry (R2: 0.98). Incorporating an understory stratification step was anticipated to increase the accuracy of chemistry estimates; however, R2 values were unchanged. While plot data suggested AISA chemistry prediction performed well, significant bidirectional reflectance effects were evident; this effect was absent in the AVIRIS data.

In the current study, 2002 AVIRIS and 2006 AISA high resolution imagery were applied to the spectroscopic investigation of spatio-temporal variations in forest chemistry. Focused primarily on the foliar biochemistry of Douglas-fir (Pseudotsuga menziesii) stands within the Greater Victoria Watershed, Victoria, BC, Canada, samples were collected and relationships between chemistry and reflectance were established. Partial least-squares regression (PLS) was employed to estimate chemistry from canopy reflectance spectra. Data were stratified through the use of hyperspectral/LIDAR forest products, transformed to second order derivative imagery and chemistry maps generated from the PLS coefficients. Chemistry estimation achieved R 2 values ≥ 0.93 for both datasets. The PLS models applied to the hyperspectral imagery yielded two temporally discrete chemistry maps. Temporal and spatial differences were investigated. Conditions, anticipated to significantly impact forest chemistry, exhibited correspondence with spatial variations in the new forest information chemistry product.

Monitoring of the 418 million ha of forests in Canada is needed to ensure the sustainable development of these forests. Hyperspectral sensing can provide mapping of forest species, forest health, and above-ground biomass. Airborne two-meter AISA hyperspectral and LIDAR data were acquired by the University of Victoria (UVic) over the Greater Victoria Watershed District (GVWD) test site and compared to NASA's AVIRIS data that had been acquired in the summer of 2002 at 4 m spatial resolution. Tree heights derived from LIDAR data, and allometric equations were used to provide independent ground estimates of biomass. Between-sensor calibration calibrated the AISA data to the same basis as the AVIRIS data. The calibrated reflectance data were used to generate forest species classifications, and biomass estimates for the test site. Average classification accuracies exceeded 89% in mapping major forest species. These products were used to create a map of above-ground carbon for the forested portion of the GVWD test site.