Creating a Drought Prediction Model Using Convolutional Neural Networks
AJAS · 2025 Earth and Environmental Sciences (inferred)
Overview
Droughts kill over 45,000 people yearly and affect the livelihoods of 55 million others, with climate change likely to worsen these effects. Despite this, researchers have struggled to develop a method that accurately predicts the location of droughts. One of the most recent and accurate drought prediction models is DroughtCast. DroughtCast utilized a Neural Network along with a plethora of weather data to predict the United States Drought Monitor (USDM) index of a given week; however, this model did not consider the contextual aspect of weather forecasting. Predicting weather exclusively using data from a single point will never be as successful as predicting the weather using data from that point and its surroundings. As a result, the researcher created a novel Convolutional Neural Network (CNN) based upon the U-Net architecture to predict future USDM indices by using the current USDM index and weather data. A 10- year (2010-2019) dataset containing a multitude of weather data, such as precipitation and Snow Water Equivalent, which was obtained from DAYMET - a NASA database for weather across all of North America. The model was trained using data from 2010, 2013, 2014, 2016, 2018, and 2019. The model was tested using data from 2011, 2012, 2015, and 2017. In comparison to DroughtCast, the Mean-Squared-Error of the CNN Model dropped by 85%, 98%, and 97% for prediction times of 1 week, 6 weeks, and 12 weeks respectively.
Competition history
- AJAS 2025
Related projects
CSEF · 2019
Predicting Wildfires with Neural Networks: An Approach to Preventing California Wildfires
JSHS · 2022
NEW YORK – LONG ISLAND Utilizing Short-Term Memory (LSTM) Machine Learning Algorithm to Create Soil Moisture Prediction Models and Improve Water Productivity in Southern California
ISEF · 2025
SoyCast: A Climate-Driven Deep Neural Network to Forecast County-Level Soybean Yields and Identify Environmentally Optimal Planting Regions
CSEF · 2019
Constructing Vegetation-Health-Map Image Forecasts Using a Novel Variable Length Attention ConvLSTM Network
AJAS · 2026
Using a Deep Learning Model to Predict Flash Flood Likelihood
AJAS · 2022
Novel Model Created That Can Predict Wildfire Burn Area Based on the Drought Factor
ISEF · 2024
Developing a Novel Wildfire Spread Prediction Model Using Spatio-Temporal Analysis and Time-series Forecasting
ISEF · 2019
The Effect of Atmospheric Conditions on Flash Flood Prediction Using Deep Learning
Closest projects by meaning, across every fair and year in the corpus.
Browse more like this
Source: AAAS Annual Meeting (Confex) / American Junior Academy of Science