Geospatially mapping carbon stock for mountainous forest classes using InVEST model and Sentinel-2 data: a case of Bagrote valley in the Karakoram range

Abstract

Geospatially assessing carbon storage is an emerging field inland science and landscape ecology, particularly in the context of the provision of ecosystem services such as global climate regulation. However, modeling the spatial distributions of carbon stocks in the heterogeneous mountainous ecosystems always remained challenging due to fragile, inaccessible, and rough terrains. This paper aims to assess the status and spatial variations in the net amount and economic value of carbon stocked by the different forest land use/land cover (LULC) classes in Bagrote Valley, Gilgit-Baltistan. To achieve this, we used the InVEST carbon and sequestration model with the LULC types identified through the object-based classification of the Sentinel-2 imagery. The results show that the carbon mainly stocked in 6 forest types out of a total of 14 LULC classes. The total carbon stock in the valley is 8114.10 tonnes in 2016, for which the dense conifer forests are the major contributor with 5409.16 tonnes of carbon. The economic value of carbon storage is approximately 74728US$/ha in the study area. The developed methodology will help to measure carbon storage in different forest types and their effects on ecosystem services at the landscape level. The research output helps policymakers to simulate the ecosystem carbon storage and its trade-offs for various environmental and economic goals.