Search results

The University of Michigan (U-M) is planning its course toward carbon neutrality. A key component in U-M carbon accounting is the calculation of carbon sinks via estimation of carbon storage and biosequestration on U-M landholdings. Here, this paper aims to compare multiple remote sensing methods across U-M natural lands and urban campuses to determine the accurate and efficient protocol for land assessment and ecosystem service valuation that other institutions may scale as relevant.

This paper tested three remote sensing methods to determine land use and land cover (LULC), namely, unsupervised classification, supervised classification and supervised classification incorporating delineated wetlands. Using confusion matrices, this paper tested remote sensing approaches to ground-truthed data, the paper obtained via field-based vegetation surveys across a subset of U-M landholdings.

In natural areas, supervised classification incorporating delineated wetlands was the most accurate and efficient approach. In urban settings, maps incorporating institutional knowledge and campus tree surveys better estimated LULC. Using LULC and literature-based carbon data, this paper estimated that U-M lands store 1.37–3.68 million metric tons of carbon and sequester 45,000–86,000 Mt CO₂e/yr, valued at $2.2m–$4.3m annually ($50/metric ton, social cost of carbon).

This paper compared methods to identify an efficient and accurate remote sensing methodology to identify LULC and estimate carbon storage, biosequestration rates and economic values of ecosystem services provided.