Many plant water use models predict leaves maximize carbon assimilation while minimizing water loss via transpiration. Alternate scenarios may occur at high temperature, including heat avoidance, where leaves increase water loss to evaporatively cool regardless of carbon uptake; or heat failure, where leaves non-adaptively lose water also regardless of carbon uptake. We hypothesized that these alternative scenarios are common in species exposed to hot environments, with heat avoidance more common in species with high construction-cost leaves. Diurnal measurements of leaf gas exchange and temperature for 11 Sonoran Desert species revealed 37% of these species increased transpiration in the absence of increased carbon uptake. High leaf mass per area partially predicted this behavior (r²=0.39). These data are consistent with heat avoidance and heat failure, but failure is less likely given the ecological dominance of the focal species. These behaviors are not yet captured in any extant plant water use model.

Field measurements occurred in 2018 between 4 June and 6 July in natural habitat at the Desert Botanical Garden (DBG) in Phoenix, Arizona, USA (33.46° N, 111.94° W, elevation 330 m). DBG is located in the northern Sonoran Desert and is characterized by hot and dry summers (June mean maximum Tair = 44 °C; average relative humidity (RH) = 18%), with mean annual precipitation (MAP) totaling 191 mm yr^-1, 35% of which falls during the summer monsoon.

We selected 11 common Sonoran Desert woody plant species (three individuals per species = 33 plants). Selection was based on diversity of leaf and canopy architecture. Selected individuals were similar in size within species, fully exposed to sunlight, and occurred on similar terrain. All species were C3 with no known CAM expression, in order to focus on species that do not carry out nighttime gas exchange.

Species measured and collected: Chilopsis linearis (Cav.); Dodonaea viscosa (L.) Jacq.; Encelia farinosa A. Gray ex Torr.; Larrea tridentata (DC.) Coville; Olneya tesota A. Gray.; Populus fremontii S. Watson; Prosopis velutina (Wooton) Sarg.; Quercus turbinella Greene; Rhus ovata S. Watson; Simmondsia chinensis (Link) C.K. Schneid.; Vauquelinia californica (Torr.) Sarg.

Data collected: leaf gas exchange (net photosynthesis, transpiration and stomatal conductance); leaf traits (leaf area, leaf mass (dry and fresh)); leaf hydraulic traits (predawn and midday leaf water potential); microclimate (air temperature, relative humidity, vapor pressure deficit, wind speed); leaf temperature; soil temperature and moisture.

Data processing was conducted using R software and manually using Microsoft Office Excel.

Each data file has a respective metadata file describing the data collected, units, and how sampling and collection occurred. Metadata files explained missing values (i.e., NA) when available. Data are uploaded as raw, processed, and/or filtered.

Data are uploaded as *.csv (one data file, one metadata file per dataset) and as *.xlsx (one single file with a tab for data and another for metadata per dataset).