Controls on delta formation, area, and topset slope: New predictive relationships developed using a global delta dataset

River deltas are irreplaceable natural and societal resources, though they are at risk of drowning due to sea-level rise and decreased sediment delivery. To enhance hazard mitigation efforts in the face of global environmental change, we must understand the controls on delta growth. Previous empirical studies of delta growth are based on small datasets and often biased towards large, river-dominated deltas. We are currently lacking relationships that predict delta formation, area, or topset slope across the full breadth of global deltas. To this end, we developed a global dataset of 5,229 rivers (with and without deltas) paired with nine upstream (e.g., sediment discharge) and four downstream (e.g., wave height) environmental variables. Using Google Earth imagery, we identify all coastal river mouths (≥ 50 m wide) connected to an upstream catchment, and define deltas as river mouths that split into two or more distributary channels, end in a depositional protrusion from the shoreline, or do both. Delta area is defined as the area of the polygon connecting the delta node, two lateral shoreline extent points, and the basinward-most extent of the delta. Topset slope is calculated as the average, linear slope from the delta node elevation (extracted from SRTM data) to the main channel mouth, and shoreline and basinward extent points. Of the 5,229 rivers in our dataset, 1,816 (35%) have a delta. Using 495 rivers (those with data available for all variables), we build an empirically-derived relationship that predicts delta formation with 76% success. Delta formation is controlled predominantly by upstream water and sediment discharge, with secondary control by downstream waves and tides that suppress delta formation. For those rivers that do form deltas, we show that delta area is best predicted by sediment discharge, bathymetric slope, and drainage basin area (R2 = 0.95, n = 170), and exhibits a negative power-law relationship with topset slope (R2 = 0.85, n = 1,342). Topset slope is best predicted by grain size and wave height (R2 = 0.50, n = 358). These empirical relationships can aid in forecasting delta response to continued global environmental change.