Assessing and modeling the impact of land use and changes in land cover related to carbon storage in a western basin in Mexico

In this study, we analyzed the land use and land cover (LULC) change in a hydrologic basin in the western coast of central Mexico and its regional variation in carbon storage. Three thematic maps (1986, 2001 and 2017) were produced by using pixel-based unsupervised classification techniques of Landsat images. LULC maps for 2017, 2033 and 2050 were modeled using a Cellular Automata Markov chain, validating their predictive power using Kappa variations. The InVEST software was used to estimate the carbon stored in four reservoirs, analyzing their variations over time (1986-2050). Accuracy assessment for the classifications revealed satisfactory results with an Overall accuracy of 83% and a Kappa coefficient of 0.76. Results show that the main landscape modifier was the exposed soils class increasing by 65% its extent in 1986, with a net increase around 87,900 ha. The evergreen forest (EF) and the tropical dry forest (TDF) classes showed a decrease along the period analyzed, displaying a net loss of 57,200 ha and 47,200 ha, respectively. Projected land cover changes followed the same trend observed, with a decreasing tendency in the EF and TDF coverages as consequence of exposed soil class increase. This change implies a reduction in the estimated total carbon stock, decreasing from 362.9 Tg C in 1986, to 336.2 Tg C in 2017. According to our model, if the trend detected in this analysis continues, it is expected a reduction to 317.9 Tg C in 2050. This approach combining Cellular Automata and Markov chain analysis with the InVEST model provides elements to estimate changes on carbon storage resulting from landscape changes. It also allows the identification of potential changes in the selected land use classes, providing with technical elements that could help to decision makers to better maintain the ecological integrity of the drainage basin.