MANGROVE CLASSIFICATION OF INDUS DELTA USING SATELLITE IMAGES

Project Title

MANGROVE CLASSIFICATION OF INDUS DELTA USING SATELLITE IMAGES

Project Area of Specialization Computer ScienceProject Summary

This study aims to provide the comprehensive mangrove species classification across the Indus river delta in Pakistan. The Indus delta extends over an area of some 6000 hectares, on the border between Pakistan and India. The mangroves in this region are the sixth largest in the world. The Google Earth Engine (GEE) a geospatial processing service powered by Google Cloud Platform, will be used, Different spervised classification ML algorithms will be applied on acquired Satellite images to classify major covers: ‘mangroves’, ‘water’ and ‘other’. After that, ‘mangroves’ will be further classified into species. According to Flora of Pakistan, eight species of mangroves have been reported along the coast of Pakistan, out of which four species have completely disappeared, three species, red mangrove (Rhizophora mucronata), black mangrove (Aegiceras corniculatum) and Indian mangrove (Ceriops tagal) are at the verge of extinction and only one species, grey mangrove (Alvicennia marina) is surviving in Indus delta. This study-project will classify the remaining four species that are thriving in this area and map their explicit locations.

Project Objectives

The current study is designed to achieve the following objective to assess the mangrove conservation and sustainability over the Indus delta in Pakistan.

• Classification of mangrove species using satellite images, across the Indus delta in Pakistan.
• Change detection in mangrove species across the Indus delta.

Project Implementation Method

This study aims to examine the changes in mangrove species using the approaches of supervised classification, the technique of training a model using some training data and apply the results/logic on the remaining/test data.

The flow of the project would be as following.

1. Collection of year wise satellite images from landsat or sentinel sensors.
2. Processing those images for cloud free experience and maximum spatial resolution.
3. Collection of data i.e. coordinates for the different mangrove species.
4. Supervised classification i.e. classifying data into training and testing data.
5. Implementation and evaluation of different machine learning algorithms.
6. Evaluations of accuracy provided by those algorithms.
7. Model creations and prediction of mangroves.
8. Oberving the changes in different species of mangroves.
9. Calculating the area of the mangrove forests across Indus delta.
10. Documenting those changes in a research.

Benefits of the Project

Understanding mangrove ecosystems and mapping their extent is critical to meeting UN sustainable developement goals (UN sustainable development indicator 6.6.1: Change in the extent of water-related ecosystems over time.). Mangroves are critical ecosystems, provide coastal protection from storm surges, maintain our climate, control floods, and stabilize coastlines. Additionally, they serve as nurseries for a number of marine wildlife species. Mangroves are also integral to the blue carbon family. Blue carbon is carbon stored, sequestered, or released from coastal vegetation ecosystems (Heer et al. 2012). Therefore, understanding mangrove extent and biomass is essential to managing the sustainability of our water ecosystems.

Remote Sensing applications are the core benefits of this project. Without doing enormous amount of field work and calculating the changes physically, we can use satellite images and power of machine learning to observe the changes in mangrove forests.

This study would also be beneficial in understanding the mangrove science and can be used to detect the change in different environmental factors with change in extent of different mangrove species.

Technical Details of Final Deliverable

The final deliverable for this project would be:

• Map mangrove extent (Indus delta specific).
• Perform time series analysis of mangrove ecosystems.
• Calculate mangrove extent in hectares.
• An app to display data using results from Indus delta.

Final Deliverable of the Project Software SystemCore Industry ITOther Industries Agriculture , Others Core Technology OthersOther Technologies Artificial Intelligence(AI), Big DataSustainable Development Goals Clean Water and Sanitation, Climate ActionRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	80000
Google cloud platform subscription.	Equipment	1	30000	30000
Traveling cost	Miscellaneous	2	5000	10000
GEE API	Equipment	1	16000	16000
Research Publication registration	Equipment	1	24000	24000