1. Climate temporality is a phenomenon that affects species' activity and distribution patterns across spatial and temporal scales. Despite the global availability of microclimatic data, their use to predict activity patterns and distributions remains scarce, particularly at fine temporal scales (e.g., < month). Predicting activity patterns based on climatic data may allow us to foresee some of the consequences of climate change, particularly for ectothermic vertebrates.

2. The Gila monster exhibits marked daily and seasonal activity patterns linked to physiology and reproduction. Here we evaluate if ecological niche models fitted using microclimate data can predict temporal activity patterns using the Gila monster (Heloderma suspectum) as a study system. Further, we identified if the activity patterns are related to physiological constraints.

3. We used dated occurrences from museum specimens and human observations to generate and test ecological niche models using minimum-volume ellipsoids. We generated hourly microclimatic data for each occurrence site for ten years using the NicheMapR package. For ecological niche modeling, we compared the traditional seasonal approach versus a daily activity pattern strategy for model construction. We tested both using the omission rate of independent observations (citizen science data). Finally, we tested if unimodal and bimodal activity patterns for each season could be recreated through ecological niche modeling and if these patterns followed known physiological constraints.

4. The unimodal and bimodal activity patterns previously reported directly from tracking individuals across the year were recovered by using niche modeling and microclimate across the species' geographical range. We found that upper thermal tolerances can explain the daily activity patterns of this species.

5. We conclude that ecological niche models trained with microclimatic data can be used to predict activity patterns at fine temporal scales, particularly on ectotherm species of arid zones coping with rapid climate modifications. Further, the use of fine temporal scale variables can lead to a better niche delimitation, enhancing the results of any research objective that uses correlative models.

We generated an ecological niche model using georeferenced specimen records (see Occurrence data) and microclimatic data at an hourly resolution (see Environmental data). Since the specimen records did not include the exact hour of collection, we made use of an available dataset of daily activity patterns for every season throughout the year based on individual observations and radio-tracked individuals of Heloderma suspectum (see daily activity pattern section below, Beck 2005). We associated the microclimatic data that corresponded to the time of activity sensu Beck (2005) according to the specific season for each record. We used the prediction of the ecological niche model for occurrence sites (suitability) to recover the daily activity patterns per season (see Daily activity patterns and microclimate variation). Finally, we validated the model's predictions regarding suitability per hour for each season concerning Beck's activity data and to iNaturalist records that had the time of sighting (see Ecological niche modeling). 

R software