Rare and extreme climatic events can leave important but difficult-to-study legacies for ecological communities. Theory suggests that the relative abundance and even presence of species in communities may be driven as much by historical contingencies associated with rare events as by filtering from average environmental conditions. Much of what we know about the consequences of extreme events comes from opportunistic studies, where ecologists were in the right place at the right time. Climate change forecasts for the 21st century include an increase in frequency of extreme events, and ecologists are increasingly called upon to predict their consequences. We must therefore exploit opportunities to quantify the ecological effects of climatic extremes when and where we can. This study examined how a rare and extreme forcing event – the unprecedented flooding associated with Hurricane Harvey – modified the composition of ant communities in east Texas. We surveyed ant communities of the Big Thicket National Preserve (BTNP) through pitfall trapping at 18 sites throughout BTNP. We conducted twice-annual sampling in 2014 and 2015, before the hurricane, and monthly sampling for one year in the aftermath of Hurricane Harvey (August 2017), beginning in November 2017. BTNP ant communities include many native species as well as two conspicuous invaders: the fire and Solenopsis invicta and the tawny crazy ant Nylanderia fulva. Thus, beyond characterizing overall trends in ant community composition, we were particularly interested in the temporal dynamics of these species' distribution and abundance throughout the Big Thicket preserve, and their responses to Hurricane Harvey. 

Sampling took place at 18 sites spanning 11 administrative units of the Big Thicket National preserve. Sampling was conducted twice per year (in May and September) for two years (2014 and 2015), and then monthly from November 2017 through November 2018 (with the exception of no sampling in December 2018). Locations for sample sites were constant across sampling events with the exception of one site that was relocated nearby (within 1 km) after it became inaccessible in Fall 2015 due to localized flooding. At each sampling site, five pitfall traps were placed at each corner of a 10 m by 10 m square with one pitfall in the center of the square (these are called "stations" and labeled A-E). Each pitfall consisted of a 50 mL centrifuge tube filled with 25 mL of reverse-osmosis water and a few drops of unscented dish soap to break the surface tension. The pitfalls were positioned such that the open lid of the tube was flush with the surface of the ground and were recovered after seven days. After each sampling period, we transferred ants into vials containing 70% ethanol and stored the vials at -20 degrees C. Workers were keyed to genus and species. 

The main data file is Pitfall_Data.csv. Each row of this data frame gives the abundance of one species at a given sampling date - site - station. If there were no ants captured for a given date - site - station then the "Ants?" field is given a zero. When multiple species were captured for a given date - site - station these are reported in separate rows. Site numbers correspond to those in the Site_Characteristics.csv data file. 

The Site_Data.csv file gives additional information about cases of missing observations in the pitfall data. This data frame indicates, for each date - site - station, whether the site was accessible for set-up, whether it was accessible to re-visit, and whether pitfalls were retrieved. This data frame also provides measurements of volumetric water content (VWC), which were collected in 2018 but not previous years. 

All metadata are described in the ReadMe file.