Unravelling indoor temperature response to summer heat through long‐term crowdsourced observations in Dutch residences

Abstract Climate change is projected to enhance the frequency of heat events, also raising indoor heat loads for urban dwellers. However, understanding of the relationship between outdoor and indoor heat remains limited. This study uses a unique indoor temperature dataset collected by citizen weather stations in seven residences (spanning 2–27 years) to observe, analyse, and model summertime indoor air‐temperature evolution in the Netherlands. We find that indoor temperature typically warms and cools more slowly than outdoor temperature, with an average lag of 260 minutes in the diurnal cycle. Furthermore, heatwave signals can persist indoors for about five days after a heatwave. To identify the dominant physical processes governing the daily indoor temperature evolution, we use a physics‐informed statistical model. This model incorporates outdoor–indoor heat convection, building conduction, and solar and thermal radiation. Model results across the seven residences are promising, with a root‐mean‐square error and median absolute error of and , respectively. A sensitivity experiment indicated that an observation period of approximately three months is required to train and test a reasonably accurate model, found by evaluating the model performance by varying the training size (0.5%–70%) and keeping the testing size constant (30%). Sensitivity experiments indicate that the model performs similarly for cooling and warming phases, though individual houses may show variation. In addition, model performance is stronger for mean temperature evolution than for minimum and maximum temperature evolution, likely due to smaller signal variability. This study demonstrates the potential of combining physics‐informed modelling with crowdsourced data to advance analyses on indoor temperature dynamics during summer in the Netherlands.