Hydroperiod, or the amount of time a lentic waterbody contains water, shapes communities of aquatic organisms. Precise measurement of hydroperiod features such as inundation timing and duration can help predict community dynamics and ecosystem stability. In areas defined by high spatial and temporal variability, fine-scale temporal variation in inundation timing and duration may drive community structure, but that variation may not be captured using common approaches including remote sensing technology. Here, we provide methods to accurately capture inundation timing by fitting hidden Markov models to measurements of daily temperature standard deviation collected from temperature loggers. We describe a rugged housing design to protect loggers from physical damage and apply our methods to a group of intermittent ponds in southeastern Arizona, showing that initial pond inundation timing is highly variable across a small geographic scale (~50km2). We also compare a 1-logger (pond only) and 2-logger (pond + control) design and show that, although a single logger may be sufficient to capture inundation timing in most cases, a 2-logger design can increase confidence in results. These methods are cost-effective and show promise in capturing variation in intraregional inundation timing that may have profound effects on aquatic communities, with implications for how these communities may respond to hydroperiod alteration from a changing climate.