FireCast: Leveraging Deep Learning to Predict Wildfire Spread
Authors: David Radke, Anna Hessler, Dan Ellsworth
IJCAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We present and evaluate a novel system, Fire Cast. Fire Cast combines artificial intelligence (AI) techniques with data collection strategies from geographic information systems (GIS). Fire Cast predicts which areas surrounding a burning wildfire have high-risk of near-future wildfire spread, based on historical fire data and using modest computational resources. Fire Cast is compared to a random prediction model and a commonly used wildfire spread model, Farsite, outperforming both with respect to total accuracy, recall, and F-score. |
| Researcher Affiliation | Academia | David Radke1 , Anna Hessler2 and Dan Ellsworth2 1David R. Cheriton School of Computer Science, University of Waterloo 2Department of Mathematics and Computer Science, Colorado College |
| Pseudocode | No | The paper provides a table detailing the model architecture (Table 1) and describes its components, but it does not include pseudocode or clearly labeled algorithm blocks with structured steps. |
| Open Source Code | No | The paper does not include an unambiguous statement that the authors are releasing their code, nor does it provide a direct link to a source-code repository. |
| Open Datasets | Yes | Our evaluation uses historical fire perimeters from Geo MAC, a United States Geological Survey (USGS) database, for each of the training and testing fires2. The Geo MAC database contains perimeter data for a variety of fires... Landsat8 satellite imagery is used as a visual input to the model, collected from Glo Vis3, another USGS database. ... A 30-meter resolution DEM is obtained from the USGS National Map4 for each fire location and geo-rectified with the Landsat8 pixels. ... the most precise and accurate historical atmospheric data available for each remote location was collected from the National Oceanic and Atmospheric Administration (NOAA)5 |
| Dataset Splits | Yes | For each fire in the training dataset, referred to as training fires, the model is exposed to an initial fire perimeter and the future 24 hours of weather data. The ground truth fire perimeter for the following day is used as validation data during training. |
| Hardware Specification | No | The paper does not specify any particular hardware used for running the experiments, such as specific GPU or CPU models. |
| Software Dependencies | No | The paper mentions that the CNN is 'implemented using Keras with the Tensor Flow backend.' However, it does not provide specific version numbers for Keras or TensorFlow, which is necessary for reproducibility. |
| Experiment Setup | Yes | The full CNN is composed of two convolutional layers of 32 and 64 hidden nodes, and uses Sigmoid and Re LU activation functions respectively. ... Both the CNN and the final dense layer use the binary crossentropy loss function and do not use regularizers. The CNN uses an RMSProp optimizer and the final dense layer uses a stochastic gradient decent optimizer. |