Using Long Short-Term Memory (LSTM) and Internet of Things (IoT) for localized surface temperature forecasting in an urban environment

Abstract:

The rising temperature is one of the key indicators of a warming climate. Its impact on urban areas becomes more severe due to the urban heat island effect compared to other landscapes. Urban areas typically have a higher temperature than the surrounding rural areas. It poses increasing risks of heat-related and illnesses as urban areas are usually populated. Therefore, it is essential to adequately monitor the local temperature dynamics to mitigate risks associated with increasing global temperatures. Since local temperature profiles, at a street level, tend to be quite different from regional weather forecasts, predictions for local temperatures directly adopted from regional weather forecasts are limited, although regional weather forecasts usually have adequate spatio-temporal coverage. There remains a clear need for an accurate air temperature prediction at the sub-urban scale on an hourly basis.

This research proposed a framework by integrating long-term historical in-situ observations and Internet of Things (IoT) observations together to train a Long Short-Term Memory (LSTM) network for air temperature prediction within the city of New York. By leveraging the historical air temperature data from in-situ observations, the LSTM model can be exposed to more historical patterns that might not be present in the IoT observations. Meanwhile, by using IoT observations, the spatial resolution of air temperature predictions is significantly improved.