The State Impact Tables dataset combines Drought Impact Reporter (DIR) data with U.S. Drought Monitor (USDM) data, joined according to the start date of each impact. Providing dates for impacts is complex and sometimes imprecise, but start dates tend to be more consistent than end dates.
DIR data is filtered to exclude impacts starting Jan. 1, as they tend to be referring to aggregated data, such as the number of wildfires during the calendar year, rather than the actual occurrence date of specific wildfires.
Impacts affecting 10 or more states are filtered out, as they may be more macroeconomic or too general for a state table. Impacts with nine or fewer states are included as they may be helpful for characterizing impacts in regions such as the Northeast that have smaller states.
Then the data are filtered to include only one impact per drought level per state. Multiples occur when a single impact affects many different locations. For DIR purposes, each state, county or city is considered a separate location, so impacts typically include multiple locations. Users may notice that the description field for older impacts include less source citation information. DIR data collection practices evolved over time. More complete information on impacts is available via the DIR. To find detail for a specific impact, use the Impact ID filter.
The interactive state tables build on conceptual state tables introduced in Noel et al. (2020), which used DIR data subset to a single drought onset period and human curation to create lists of impacts that occurred at each drought level for each state. This replaced an older, hypothetical list of what impacts might be expected at each level of the USDM, with one list for the entire country.
Response to the original state tables was positive, but with requests for more complete and updated data. This led to creation of the interactive version, first launched in 2022. The dataset for the interactive State Drought Impacts tool is created with an R script, less labor-intensive to update than the human-curated tables. A possible downside is that the lists of impacts are much longer, but the filters help. Supervised AI may also be useful for summarizing lists of impacts.