Rethinking heat in thousands of school classrooms through continuous monitoring and novel exposure metrics

As global temperatures rise, heat exposure in classrooms is becoming a growing concern for the millions of students attending school, in particular those attending buildings without air conditioning (AC). With limited resources and competing interests, school decision-makers need health-related data-based approaches to inform cooling solutions and prioritize investments. In collaboration with a large school district in United States (US), we analyzed minute-level temperature data in >3,600 classrooms across 125 schools during the 2023 hot season. Using a first-of-its-kind commercial-grade indoor sensor network and data science methods, we quantified heat exposure through novel heat metrics capturing intensity, frequency, and duration, and characterized spatial variability within and across buildings with three types of AC. On average, intra-building temperature variability was 2.3 degrees Celsius (°C), with a maximum value of 14.3°C. On a hot day classrooms exceeded extreme caution thresholds by 0.1%, 1.1%, and 8.4% in schools with central, window, and no AC, respectively. Classrooms on the top floor were 0.3°C, 0.5°C, and 5.7°C warmer than classrooms on the first floor, for central, window, and no AC groups, respectively. Novel and traditional heat exposure metrics were weakly correlated, with implications for school rankings. Findings identified schools with the greatest cooling needs and investigated key predictors of classroom overheating. Our results underscore the need for continuous thermal monitoring in all classrooms and highlight the importance of access to mechanical cooling in locations that have historically not been prepared for extreme heat. Our work shows how data analyses informed through researcher-school partnerships can support critical climate resilience needs in schools. • We found considerable variability in indoor heat exposure between and within school classrooms, across different air conditioning status, in a large, urban school district in Northeast US. • Classrooms with central AC recorded the coolest conditions, specially during the hottest periods. • Classrooms at ground levels were up to ~6°C cooler than those in upper levels, with smaller differences in schools with mechanical cooling. • Novel metrics complemented central tendency thermal characterization by capturing frequency, duration, and intensity nuances of heat exposure. • Comprehensive temperature monitoring in all classrooms can help districts identify schools with the greatest cooling needs and prioritize cooling investments.