Checking for non-preferred file/folder path names (may take a long time depending on the number of files/folders) ...

Input data and code script for comparing strategies for training LSTM models for street-scale flood prediction in Norfolk, Virginia


Authors:
Owners: This resource does not have an owner who is an active HydroShare user. Contact CUAHSI (help@cuahsi.org) for information on this resource.
Type: Resource
Storage: The size of this resource is 586.0 KB
Created: Feb 06, 2026 at 4:42 a.m. (UTC)
Last updated: Feb 06, 2026 at 6:59 p.m. (UTC)
Citation: See how to cite this resource
Content types: CSV Content 
Sharing Status: Public
Views: 26
Downloads: 1
+1 Votes: Be the first one to 
 this.
Comments: No comments (yet)

Abstract

These are tabular input data and python code script for LSTM surrogate model built for real-time street flood prediction in Norfolk, VA. The LSTM surrogate model approximates water depth on streets generated by TUFLOW. The inputs of the model are topographic feature: elevation, and environmental features such as hourly rainfall, cumulative rainfall in previous hours, hourly tide level, etc. The output of the model is hourly water depth on streets during storm events generated by the TUFLOW model.

Subject Keywords

Coverage

Spatial

Coordinate System/Geographic Projection:
WGS 84 EPSG:4326
Coordinate Units:
Decimal degrees
Place/Area Name:
Norfolk, VA
North Latitude
36.9229°
East Longitude
-76.2422°
South Latitude
36.8411°
West Longitude
-76.3304°

Temporal

Start Date:
End Date:

Content

README.txt

README

The instructions for reproducing each experiment are provided below.

* 1st Experiment

- Download the file '40_flood-prone_streets.zip'.
- Download the file '1st_experiment.ipynb'.
  This notebook is currently configured for the Global Model.
- Refer to 'Clustering_result.docx' and cluster the 40 streets according to the provided clustering results.
- Modify '1st_experiment.ipynb' as needed based on the number of streets being predicted.

* 2nd Experiment

- The second experiment focuses on Cluster 4 (K = 8), which includes 6 streets.
- Download the file '2nd_experiment.ipynb'.
  This notebook is currently configured such that the water depth feature at time t−1 is excluded for all six streets.
- Run the code while varying which water depth features at time t−1 are included.

* 3rd Experiment

- Download 'FID_6508.csv' and '3rd_experiment.ipynb'.
- The current version of '3rd_experiment.ipynb' is configured to include Street 6508 in the global model.
- Modify the code to match the number of streets in the cluster to which street 6508 is assigned, and then run the experiment.

How to Cite

Jeong, J. (2026). Input data and code script for comparing strategies for training LSTM models for street-scale flood prediction in Norfolk, Virginia, HydroShare, http://www.hydroshare.org/resource/f384de0d4bb248ba915eac7103822f9d

This resource is shared under the Creative Commons Attribution CC BY.

http://creativecommons.org/licenses/by/4.0/
CC-BY

Comments

There are currently no comments

New Comment

required