This data release contains model inputs, R code, and model outputs for predicting depth to bedrock in the Delaware River Basin at a 1km gridded resolution with a random forest model. Model inputs are provided in a comma-separated value (csv) file. The training data used in this study of 72,773 point observations of depth to bedrock (DTB) within the Delaware River Basin (DRB) that was compiled from several sources. These data were attributed with 15 predictor variables representing topographic, soil, geologic, and physiographic characteristics of the depth to bedrock observation. One predictor variable is a grouped surficial geology category that was adapted from the State Geologic Map Compilation (Horton and others, 2017); the grouped lithology categories are provided in this data release as a shapefile dataset. The predictions from the random forest model are provided as a gridded geoTIFF file. Two files are provided - one for uncorrected model predictions and another for predictions that were bias-corrected using the Empirical Cumulative Distribution Matching (ECDM) approach of Belitz and Stackelberg (2021). The bias-corrected predictions are the final model predictions for use in other applications. Horton, J.D., San Juan, C.A., Stoeser, D.B., 2017. The State Geologic Map Compilation (SGMC) geodatabase of the conterminous United States (Report No. 1052), Data Series. Reston, VA. https://doi.org/10.3133/ds1052 Belitz, K., Stackelberg, P.E., 2021. Evaluation of six methods for correcting bias in estimates from ensemble tree machine learning regression models. Environmental Modelling & Software 139, 105006. https://doi.org/10.1016/j.envsoft.2021.105006