A Predictive Model for Geographic Distributions of Mangroves

TL;DR

This paper develops a highly accurate support vector regressor model to predict the global distribution of mangroves based on climate variables, highlighting their ecological importance and response to climate change.

Contribution

It introduces a predictive model linking climate variables to mangrove distribution, with the support vector regressor achieving near-perfect correlation.

Findings

Support vector regressor achieved a correlation coefficient of 0.9998.

Climate variables like temperature and sea level significantly influence mangrove distribution.

The model provides a tool for forecasting mangrove habitats under climate change scenarios.

Abstract

Climate change is an impending disaster which is of pressing concern more and more every year. Countless efforts have been made to study the long-term effects of climate change on agriculture, land resources, and biodiversity. Studies involving marine life, however, are less prevalent in the literature. Our research studies the available data on the population of mangroves (groups of shrubs or small trees living in saline coastal intertidal zones) and their correlations to climate change variables, specifically, temperature, heat content, various sea levels, and sea salinity. Mangroves are especially relevant to oceanic ecosystems because of their protective nature towards other marine life, as well as their high absorption rate of carbon dioxide, and their ability to withstand varying levels of salinity of our coasts. The change in global distribution was studied based on global distributions of the previous year, as well as ocean heat content, salinity, temperature, halosteric sea level, thermosteric sea level, and total steric sea level. The best performing predictive model was a support vector regressor, which yielded a correlation coefficient of 0.9998.