A Framework for Forecasting Indoor Temperatures for Populations at Risk from Heat Waves

Previous work has shown that heat waves as experienced indoors by residents without air conditioning are quite different from the outdoor heat waves that are forecast: indoor conditions experience thermal lag and greenhouse warming compared to outdoors. The differences can be crucial to understanding health impacts and providing suitable heat alerts. Here are presented attempts at indoor forecasting driven by outdoor conditions. An energy flow model is used to model indoor temperatures, with thermal parameters for each residence based on several months of indoor data. Once these parameters are found by regression, forecasts are constructed based on current and forecast outdoor conditions. This work shows the results of daily average temperature forecasts only; more comprehensive models for sub-daily forecasts are under development but will be outlined. The sensing system will use LoRaWAN technology via the open source Things Network community so that at-risk communities without access to internet can still have sensor connectivity.