How to Determine Water Level Differences Using the Modified Normalized Difference Water Index (MNDWI) and Satellite Imagery
Introduction
Accurately monitoring water levels in lakes, rivers, and wetlands is crucial for effective water resource management, flood prevention, and environmental conservation. Traditional methods can be labor-intensive and limited in scope, but satellite imagery offers a comprehensive solution. One of the most effective techniques for identifying water bodies and determining water level differences is the Modified Normalized Difference Water Index (MNDWI). This article will delve into how the MNDWI technique, using Sentinel-2 satellite imagery, can be utilized to monitor water levels.
What is MNDWI?
The Modified Normalized Difference Water Index (MNDWI) is an advanced remote sensing technique used to enhance the visibility of open water features while suppressing the noise from built-up land, vegetation, and soil. The MNDWI is particularly useful for distinguishing water bodies in regions with mixed land cover.
How MNDWI Works
The MNDWI utilizes two specific spectral bands from satellite imagery:
- The Green band
- The Short-Wave Infrared (SWIR) band
The formula for calculating MNDWI is:
MNDWI = (Green — SWIR) / (Green + SWIR)
Values of MNDWI greater than zero generally indicate the presence of water, while negative values typically represent non-water features.
Case Study: Monitoring Water Levels in Sebkhet Sijoumi, Tunisia.
Sebkhet Sijoumi, a prominent salt flat in Manouba, Tunisia, experiences seasonal flooding. By applying the MNDWI technique to Sentinel-2 imagery, we can accurately delineate water bodies and monitor changes over time.
Steps to Determine Water Levels Using MNDWI
- Acquiring Satellite Imagery
We used Sentinel-2 satellite data, which is freely available and offers high-resolution multispectral imagery.
2. Preprocessing the Imagery
- Atmospheric Correction: We Performed an atmospheric correction to remove any distortions caused by the atmosphere, ensuring the accuracy of the data.
- Image Calibration: Calibrate the images to standardize the pixel values for accurate analysis.
The Sentinel-2 image contains 12 spectral bands with resolutions of 10, 20, and 60 meters. In this work, two bands are used: band 3 (Green) and band 11 (SWIR).
PanSharpening: The difference in resolution between band 3 (10m) and band 11 (20m) results in the need to increase the resolution of the latter. According to Du et al., the best technique for enhancing the spatial resolution of Sentinel images is the HPF (High Pass Filter) technique, as determined by comparing four different techniques. Using the “Erdas Imagine” software; under the Raster Menu and in the resolution section, you can access the “HPF Resolution Merge” tool.
3. Calculating MNDWI
After extracting and calibrating the two necessary bands for the next operation, and to apply the MNDWI formula, we adopted a workflow in the Model builder in the ERDAS Imagine software.
4. Classifying Water Bodies:
The images obtained at the end of the MNDWI process have pixel values ranging between [-1,1]. According to Du et al. (2011), water always has a positive value, with a threshold value between [0,0.4]. Based on histograms and visual identification of pixels, the threshold value is determined to be 0. Using ArcGIS, the reclassify function divides the image into two classes: “Water” and “Non-water.” Converting the “Water” class from raster format to vector format and using the vector format boundary of the study area allowed us to achieve the desired result.
Results
Conclusion
The MNDWI technique is a powerful tool for monitoring water levels using satellite imagery. Its ability to accurately identify water bodies amidst mixed land cover makes it invaluable for environmental monitoring, flood prevention, and water resource management. As satellite technology advances, techniques like MNDWI will continue to enhance our ability to manage and protect vital water resources.
References
Xu, H.Q. Modification of normalized difference water index (NDWI) to enhance open water features in remotely sensed imagery. Int. J. Remote Sens. 2006, 27, 3025–3033.
Du, Z.Q.; Li, W.B.; Zhou, D.B.; Tian, L.Q.; Ling, F.; Wang, H.L.; Gui, Y.M.; Sun, B.Y. Analysis of Landsat-8 OLI imagery for land surface water mapping. Remote Sens. Lett. 2014, 5, 672–681.