Using machine learning techniques to quantify the environmental impact of African forest elephants: combining real world conservation with informatics in Google Earth Engine

African forest elephants (Loxodonta africana cyclotis) face severe threats from poaching and habitat loss, with estimated population losses between 62-81% in central African since the early 2000s. Forest elephants are known to transport nutrients across gradients, disperse seeds, and inflict damage to the understory through browsing. It is estimated that changes in vegetation structure from elephant disturbance significantly affect carbon stocks in the Afrotropics, however these findings have not been validated with structural data. This study aimed to quantify the role forest elephants play as ecosystem engineers and their impact on habitat heterogeneity using lidar (light detection and ranging) and machine learning techniques in Google Earth Engine. Utilizing elephant dung count data, bio-logged tracking data, and environmental predictors such as precipitation, temperature, slope, and EVI from Google Earth Engine datasets, species distribution models were created using two machine learning techniques: random forest and maxent. Multiple types of lidar data collected in Lope National Park, Gabon were then compared in elephant high and low areas of the forest to determine their impact on canopy composition and structure. Taking the difference in the mean waveforms for high and low elephant areas of the park resulted in a significant difference in the low to mid canopy (5-20 m), indicating a potential elephant signal of an open understory. If elephants impact tropical forests by clearing out the understory and allowing larger trees which store more carbon to succeed, this information could be used by REDD+ to provide additional carbon credits to countries investing in elephant protection. This project would not only assist in conservation efforts of the endangered African forest elephant, but would also improve our understanding of their unique role in the ecosystem.