The Daedalia Planum is one of the volcanic plains on Mars, located near Arsia Mons volcano. Daedalia Planum has multiple lava flows, and studying them can give new insights into Mars Amazonian volcanism and the volcanic activity of Arsia Mons. In this work, I performed a characterization and classification of lava flows in the eastern region of Daedalia Planum using remote sensing data from the Thermal Emission Imaging System (THEMIS) and the Mars Reconnaissance Orbiter Context Camera (CTX). Seven distinct lava flow types, along with three additional surface classes, were identified based on differences in surface texture, dust cover, and thermal inertia.

To map these classes, classification was performed using THEMIS infrared dataset and three supervised machine learning algorithms commonly applied in Earth-based studies: Support Vector Machines (SVM), k-Nearest Neighbours (KNN), and Random Forest (RF). Random Forest achieved the highest accuracy and was subsequently applied to multiple dataset: (1) THEMIS IR bands 2-9, (2) THEMIS IR full dataset (bands, indices, thermal inertia, and texture features), (3) CTX dataset with its texture features, and (4) combined CTX and THEMIS IR datasets. Random Forest Classification run on combined THEMIS IR and CTX dataset achieved the highest classification accuracy of 96.4%.

This study demonstrates that using machine learning classifiers widely used in Earth based studies to Martian datasets provides accurate and reliable results, showing their value for future planetary volcanology studies.