ArcGIS Developers
Dashboard

ArcGIS API for Python

River Turbidity Estimation using Sentinel-2 data

Introduction

Turbidity represents the level of suspended sediments in water also indicating water clarity or how clear is the water. It is mainly caused by the presence of silt, algae in a water body, or industrial waste disposed in the rivers by mining activity, industrial operations, logging, etc.

Traditionally, turbidity is analyzed by evaluating water samples taken during field measurements. However, field studies are expensive, time and labor intensive, besides, during lockdown field surveys cannot be undertaken. Thus, a good alternative to field survey measurements is satellite remote sensing data, which can capture both spatial and temporal variations in river turbidity levels. Accordingly, Sentinel-2 multispectral data is used in the current study to evaluate the changes in river turbidity during COVID-19 lockdown, near the holy city of Allahabad, India.

Allahabad, situated in the northern part of India, is at the confluence of two major rivers: Ganga and its tributary river Yamuna. The point of confluence is a sacred place for Hindus where thousands of devotees gather everyday by the banks of the river to offer prayers. Millions of people depend on the waters from these rivers yet both rivers are heavily polluted by industrial waste, by sewage and even by the remains of the many bodies cremated on their banks. On March 25, 2020, India announced a lockdown for controlling the spread of COVID-19 leading to complete shutdown of industries and restricted human movement. The lockdown resulted in reduction of turbidity thereby improving water quality in rivers throughout the country. This notebook will elaborate the steps to measure this change in turbidity using ArcGIS API for Python.

Necessary Imports

In [1]:
import pandas as pd
from ipywidgets import HBox, VBox, Label, Layout


import arcgis
from arcgis.features import GeoAccessor, GeoSeriesAccessor
from arcgis.raster.analytics import convert_feature_to_raster, convert_raster_to_feature
from arcgis.raster.functions import extract_band, greater_than, clip, remap, colormap, stretch
from arcgis.features.analysis import dissolve_boundaries

Connect to your GIS

In [2]:
from arcgis import GIS
agol_gis = GIS('Home')
ent_gis = GIS('https://pythonapi.playground.esri.com/portal', 'arcgis_python', 'amazing_arcgis_123')

Get the data for analysis

Sentinel-2 Views is used in the analysis, this multitemporal layer consists 13 bands with 10, 20, and 60m spatial resolution. The imagery layer is rendered on-the-fly and available for visualization and analytics. This imagery layer pulls directly from the Sentinel-2 on AWS collection and is updated daily with new imagery.

In [3]:
s2 = agol_gis.content.get('fd61b9e0c69c4e14bebd50a9a968348c')
sentinel = s2.layers[0]
s2
Out[3]:
Sentinel-2 Views
Sentinel-2, 10m Multispectral, Multitemporal, 13-band images with visual renderings and indices. This Imagery Layer is sourced from the Sentinel-2 on AWS collections and is updated daily with new imagery. This layer is in beta release. Imagery Layer by esri
Last Modified: May 28, 2020
16 comments, 963,643 views

A feature layer was published which represents the extent of study area. This feature will be used for extracting Sentinel-2 tiles of study area.

In [4]:
aoi1 = agol_gis.content.search('title:allahabad_aoi owner:api_data_owner', 'Feature Layer Collection')[0]
aoi1
Out[4]:
allahabad_aoi
allahabad_aoiFeature Layer Collection by api_data_owner
Last Modified: December 02, 2020
0 comments, 0 views

Methodology

The above diagram encapsulates the overall methodology that has been followed in the estimation of the river turbidity.

Prepare data for analysis

Sentinel-2 Views imagery layer consists data for the whole world and span different time periods. Thus, the first step is to filter out the data for the rivers in Allahabad region prior to and during the lockdown period.

Create geometry of area of interest (AOI)

The geometry of AOI is created for filtering out the Sentinel-2 tiles for the study area.

In [5]:
aoi_layer = aoi1.layers[0]
aoi_feature = aoi_layer.query(where='fid=1')
aoi_geom = aoi_feature.features[0].geometry
aoi_geom['spatialReference'] = {'wkid':3857}

Filter out sentinel-2 tiles

In [6]:
m = agol_gis.map('Allahabad, India', 11)
m