Team: 10
School: Academy For Tech & Classics
Area of Science: Earth and Space Sciences
Proposal: Drought and flooding have always been principal concerns for anyone trying to live in New Mexico. With climate change, these problems have only gotten worse. Many areas of New Mexico are either approaching or experiencing a situation where citizens have limited access to running water.
To prepare for and thus combat the effects of drought and flooding, weather forecasting is an invaluable resource. However, while we can predict precipitation with a decent level of accuracy, its impact on latent factors, such as water supply levels, is not immediately obvious and needs to be modeled to provide a specific prediction.
We plan on using machine learning techniques to create a set of models which predict the state of various variables strongly affected by precipitation. For example, we may model the depth of the Rio Grande or the Santa Fe Watershed based on surrounding precipitation. As many of these relationships will be relatively simple, often even linear, our models will initially be created through regression techniques. However, it is not unreasonable to suspect that more complex ML models (such as neural networks) may be necessary when modeling more complex situations. With both complex and simple modeling techniques, we should be able to reach a relatively high degree of accuracy by tailoring our models to the specific conditions and environment of the variable we are trying to predict.
Team Members:
Sponsoring Teacher: Jenifer Hooten