Combined scores on testing locations (see S3 Appendix for metric definition).

The rows represent the models and are organized bottom to top from highest score (worst model) to lowest score (best model). The columns represent individual locations (see Table B in S7 Appendix for the names), and the vertical, white lines separate states. The states are organized left to right from highest mean score to lowest, and within each state the locations are organized left to right by descending score.