Wildfires shape landscapes and ecosystems, affecting health and infrastructure. Understanding the complex interactions between social organization, human activity and the natural environment that drive wildfire occurrence is becoming increasingly important as changing global environmental conditions combined with the expanding human-wildland interface, are expected to increase wildfire frequency and severity. This paper examines the anthropogenic drivers of wildfire, and the relationship between the organization of human activity in time and wildfire occurrence focusing on the effects of transitions into and out of Daylight Saving Time (DST). DST transitions shift activity in relation to natural wildfire risk within a solar day, induce changes in the time allocated to wildfire-causing activities and disrupt sleep patterns. The paper estimates short and medium run effects of DST-induced changes in the temporal organization of human activity through a Regression Discontinuity Design with time as the running variable and Fixed Effects models, using data from over 1.88 million non-prescribed ignitions recorded in the contiguous US over 23 years. Estimates suggest that DST has a quantitatively and statistically significant immediate and medium-run effect on wildfire occurrence. Wildfire occurrence jumps by around 30% in the immediate aftermath of transitions into DST, adding about 98 human-caused wildfires across the contiguous US per year, while the transition's effect is detectable for 3 weeks. Transitions induce within-day temporal displacement of wildfires in a pattern compatible with the shifting of human activity mechanism, while the result cannot be attributed exclusively on disruptions in sleep patterns. Naturally arising lightning-strike wildfires do not respond to changes in civil time, while the results are robust to changes in assumptions. Results suggest that wildfire policy should account for the temporal organization of human activity.