The presence of large outflow channels on Mars shows the importance of water in shaping the surface of the planet over geologic time. Chaotic terrain has been identified as the source region for flood waters responsible for carving out many of these channels. There are still many unanswered questions regarding chaotic terrains on Mars. Using the most up to date CTX, HRSC, and MOLA coverage, DEM and TIN models were used to investigate examples of smooth-topped chaotic terrains which include: Hydraotes Chaos, a crater pair in Hydaspis Chaos, Baetis Chaos, and Candor Chaos, all south of Chryse Planitia. The findings of this study suggest that the collapse of chaotic terrains is not regionally controlled. This study also suggests that the largest chaotic terrains do not require external heat sources to form. Finally, there is evidence that chaotic terrain forming events have occurred from the Middle Noachian to the Late Hesperian/Early Amazonian. This data set contains the raw data for the Chaotic terrains studied here. Specifically, for each terrain the following files are included: Baetis • Baeties+East_Statistics_All_Mesas o Spreadsheet of all Baetis and East Chaos mesas as well as some flat areas of the channel floor. • baetis+east_volume_depth_calc o Spreadsheet of volume calculations • Mosaic and DEM composites • Image of Mesa locations • TIN of Baetis and East Chaos • Volume surfaces used in volume calculation Candor • Candor_Statistics_All_Mesas o Spreadsheet of all Candor Chaos mesas as well as some flat areas of the floor. • Candor_volume_calc o Spreadsheet of volume calculations • Mosaic and DEM composites • Image of Mesa locations • TIN of Candor Chaos Hydaspis • Hydaspis_Statistics_All_Mesas o Spreadsheet of all Hydaspis Chaos mesas as well as some flat areas of the plateau. • Hydaspis_Calculation_Volume_Polygon o Spreadsheet of volume calculations • Hydaspis_Terrace_Measurements o Spreadsheet of elevation points of terraces • Mosaic and DEM composites • Image of Mesa locations • TIN of Hydaspis Chaos • Volume surfaces used in volume calculation Hydraotes • Hydraotes_Statistics_All_Mesas o Spreadsheet of all Hydraotes mesas as well as some flat areas of the plateau. • Hydraotes_Calculation_Volume o Spreadsheet of volume calculations • Hydraotes_Terrace_Elevation_Points o Spreadsheet of elevation points of terraces • Mosaic and DEM composites • Image of Mesa locations • TIN of Hydraotes Chaos

Data files and descriptions contained here:1. model_outputs: Hydrologic model output as shapefiles. Model described in detail in Horvath et al. (2016) and adapted for Mars in Horvath & Andrews‐Hanna (2017). This numerical model is a combined surface-subsurface hydrology model, with input evaporation potential (Ep) and precipitation (P) rates from Earth analog climates provided by the North American Land Data Assimilation Systems Phase-2 observation-based climate model. An annual precipitation of 300 mm/Mars year was assumed, uniformly distributed across the model domain. The total annual precipitation reaching either the surface or subsurface hydrologic system was determined using an Earth-based empirical method dependent on the aridity index (φ=Ep/P; Budyko, 1974). This approach relates the aridity index (φ) to the measured catchment discharge, deriving an estimate of the actual evaporative loss from the catchment. Subsurface flow was modeled using a finite-difference approximation to the groundwater flow equation, controlled by a depth dependent permeability (depth averaged permeability of 1 10-13 m2) and porosity (20%) distribution, and an assumed aquifer depth (10 km) based on the megaregolith aquifer model of Hanna and Phillips (2005). Runoff was modeled using a simple linear reservoir approximation which assumes catchment storage is linearly related to runoff. Lakes were allowed to naturally form on the surface where the contribution of groundwater, surface water, and precipitation directly to the lake balanced evaporation off of the lake surface. We focused on the aridity index dependence for modeled lake areas and elevations - here we provide model outputs with aridity indices of 1.5 (subhumid) and 3.5 (semiarid) (the two main climate conditions we focused on in the main paper).2. shapefiles.zip: For morphologic analysis, mapping was performed on an orthorectified and equalized basemap constructed from 6 m/pixel resolution Context Camera (CTX) imagery (Malin et al., 2007). For each crater within our region, we mapped all observed gully networks (GN) around the interior crater rim (at approximately 1:150,000 scale), which are small-scale branching erosional features. At the termination of each GN, we recorded the elevation using High-Resolution Stereo Camera (HRSC) digital elevation models (DEMs) where available (~10–30 m vertical resolution) (Jaumann et al., 2007), and the global 463 m/pixel resolution Mars Orbiter Laser Altimeter (MOLA; Smith et al., 2001) dataset otherwise (at ~100 m vertical resolution). We selected and mapped eighteen craters, six of which were previously identified as inferred paleolakes (Fassett & Head, 2008; Goudge et al., 2015; Grotzinger et al., 2014, 2015) .3. Roseborough_GRL_CraterCountData_SI_Resubmission_v0: We used the CraterTools ArcMap add-in to map crater ejecta and floors within the Gale crater region at approximately 1:100,000 scale, allowing us to capture craters at and below the 1 km diameter minimum benchmark (Kneissl et al., 2011). Crater diameters were then exported from ArcGIS. Here we report the surface area and the corresponding crater diameters for each surface we mapped (and include the crater stats figures (showing the derived modeled age) and our crater maps).