Track and Monitor river channel migration of any study area using GIS

River channel migration भनेको के हो भने:
नदीको बहाव मार्ग (channel) समयसँगै एक ठाउँबाट अर्को ठाउँतिर सर्दै जाने प्रक्रिया हो।
नदी सधैं एउटै ठाउँमा स्थिर रहँदैन, दायाँ–बायाँ सर्छ, घुमाउरो बन्छ, पुरानो धार काट्छ, नयाँ धार बनाउँछ – यही प्रक्रियालाई river channel migration भनिन्छ।

नीरसोजा भाषामा: नदीले आफ्नो बाटो “shift” गर्छ।


                                                        Fig. Channel migration

1. River channel migration के–के कारणले हुन्छ?

मुख्य कारणहरू:

  • Bank erosion (छेउको किनारा बगाएर लैजानु)

  • Sediment deposition (एक ठाउँमा गाद–बालुवा थुप्रिनु)

  • Discharge परिवर्तन (बाढी, सुक्खा, दीर्घकालीन बहाव परिवर्तन)

  • Slope र geology (माटो/ढुंगाको किसिम, ढलान, घाटीको आकार)

  • वन विनाश, खेती, sand mining, तटबन्ध, ड्याम जस्ता मानवीय गतिविधि

यी कारणले नदीको channel वर्षौंदेखि दशकौंसम्म बिस्तारै सर्दै जान्छ।

2. River channel migration का observed patterns (ढाँचा / प्रकार)

नदीले move गर्दा थुप्रै प्रकारका ढाँचाहरू देखिन्छन्। मुख्य–मुख्य यसरी सम्झिन सजिलो हुन्छ 👇

2.1 Lateral migration (side–side सर्ने)

  • नदीको घुमाउरो भाग (meander) मा धेरै देखिन्छ।

  • Outer bank (बाहिरी किनारा) मा erosion, inner bank (भित्री किनारा) मा deposition हुन्छ।

  • नदी दायाँ वा बायाँ सर्दै जान्छ → खेत/बस्ती कटान, अर्को साइडमा नयाँ जमिन थपिनु।

👉 West Rapti, Karnali जस्ता तराईका नदीहरूमा धेरै ठोस रूपमा देखिने pattern यही हो।

2.2 Downstream migration (downvalley सर्छ)

  • नदीको channel पुरानै दिशामा तलतिर (downstream) shift हुँदै जान्छ।

  • Meander bends पनि तलतिर “घिस्रिँदै” जाने जस्तै देखिन्छ।

  • लामो समय स्केलमा हेर्दा नदीको curvature downstreamतिर सरीरहेको देखिन्छ।

2.3 Meander growth, cutoff, र oxbow lake बन्ने

Meandering नदीमा देखिने classical pattern:

  1. सुरुमा सानो घुमाउरो हुन्छ।

  2. Outer bank erosion र inner bank deposition को कारणले घुमाउरो बढ्दै–बढ्दै जान्छ।

  3. कुनै बेला बाढी वा high flow मा neck cutoff हुन्छ (घुमाउरोको घाँटी काटिन्छ) र नदी छोटो सीधा मार्ग लिन्छ।

  4. पुरानो घुमाउरो भाग oxbow lake वा abandoned channel बन्छ।

यो पनि channel migration को नै एउटा pattern हो – नदीले आफ्नो पुरानो path त्यागेर नयाँ short path रोज्छ।

2.4 Avulsion (हठात् channel change)

  • जब नदीको main channel अचानक नयाँ मार्गतिर छलाङ मारेर जान्छ, ठूलो floodplainमा नयाँ channel बनाउँछ।

  • साधारण lateral migration भन्दा धेरै dramatic हुन्छ – एकै घटनामा नदीको मुख्य धार नै सरेको जस्तै।

  • विशेष गरी low gradient, sediment–भरिएको floodplain र embankment/levee failureको सन्दर्भमा देखिन्छ।

2.5 Braided rivers मा bar–bar को shifting

Braided नदी (धेरै साना धाराहरू र बीचबीचमा बालुवाका टापुहरू हुने) मा:

  • Mid–channel bars (टापु) हर वर्ष आकार र स्थान बदल्छन्।

  • कहिले एउटै channelमा धेरै धार हुन्छन्, कहिले कम हुन्छ।

  • Channel migration pattern =

    • बार shifting

    • multiple small channels को खोलाघाटा ↔ बन्द/खुला हुने।

यो पनि planform migration नै हो, तर धेरै complex pattern देखिन्छ।

2.6 Vertical adjustment (तलतिर काट्ने वा गादले माथि उठ्ने)

यद्यपि प्रायः “migration” भन्नासाथ lateral सोचिन्छ, vertical पनि महत्त्वपूर्ण छ:

  • Incision: नदी तल–तिर काटेर bed deepen गर्छ।

  • Aggradation: धेरै sediment आउने हुँदा गाद–बालुवा थुप्रिएर river bed माथि उठ्छ।

यसले पनि channel को behaviour (flooding, bank stability, migration rate) परिवर्तन गर्छ।

3. Spatial pattern (कहाँ कति migration हुन्छ?)

Observed patterns सामान्यतया यस्तो हुन्छ:

  • Outer meander bends → सर्वाधिक bank erosion, migration rate high.

  • Inner bends/point bars → deposition, accretion (नयाँ जमिन)।

  • Confining valley, rocky banks → migration सापेक्ष कम।

  • Wide alluvial plains (तराई, Deukhuri जस्ता) → channel धेरै mobile, meander amplitude ठूलो हुन्छ।

4. Temporal pattern (समय अनुसार परिवर्तन)

दीर्घकालीन observation ले देखाउँछ:

  • High–flood years पछि channel migration rate ह्वातै बढेको हुन्छ।

  • Embankment, spur, तटबन्ध बनेपछि migration pattern localize हुन्छ –
    कतै रोकिएको, कतै अचानक धेरै बढेको (downstream/other bank).

  • दशकौंदेखि हेर्दा:

    • कतिपय reach relatively stable

    • कतिपय reach चाहिँ निरन्तर सर्दै जाने “hotspot”

1. Get Suitable Data

a) Remote sensing imagery

Use multi-date images so you can compare river position over time:

  • Landsat (30 m) – free, good for decadal-scale change (from 1980s onward).

  • Sentinel-2 (10 m) – free, better spatial detail (since ~2015).

  • High-res imagery – Google Earth, commercial (Maxar/Planet) if you need fine-scale bends/bank erosion.

  • Old topographic maps / aerial photos – useful for historical channel positions (e.g., 1960s–1990s).

b) DEM / elevation data

  • SRTM / ALOS / other DEMs to understand floodplain, bank height, valley shape.

2. Preprocess the Data in GIS

In QGIS / ArcGIS:

  1. Reproject all images to the same coordinate system (e.g., UTM zone for your area).

  2. Georeference old maps or scanned aerial photos.

  3. Clip imagery to your study reach (buffer along river centerline, like 5–10 km wide).

3. Extract River Channel for Each Time Step

You want to create a polygon or line representing the river at each date.

Option A: Manual digitizing

  • On-screen digitize the water boundary or centerline for each year/date.

  • Create layers like:

    • River_1990

    • River_2000

    • River_2010

    • River_2020

This is slower but very accurate if imagery is clear.

Option B: Semi-automatic classification

  1. Water index from satellite bands, e.g.:

    • NDWI = (Green – NIR) / (Green + NIR)

    • MNDWI using Green and SWIR

  2. Threshold NDWI/MNDWI raster to separate water vs non-water.

  3. Convert water raster → polygon → clean with smoothing, removing small patches.

Do this for each date so you have river polygons for multiple years.

4. Derive Centerlines and Bank Lines

From each river polygon:

  • Use “Polygon to Centerline” or skeletonize tools (plugins in QGIS or ArcGIS) to get:

    • Centerline (thalweg approximation).

    • Alternatively, manually trace a centerline along the channel.

  • You can also generate:

    • Left bank and right bank polylines by converting polygon boundaries and splitting at confluences.

Name them by year: Centerline_1990, Centerline_2000, etc.

5. Measure Channel Migration

Here are common methods:

a) Bankline migration distance

  1. For two years (e.g., 2000 & 2020):

    • Take right bank 2000 and right bank 2020.

  2. Create measurement transects perpendicular to the river flow:

    • Use a tool like DSAS (Digital Shoreline Analysis System) in ArcGIS, or similar scripts in QGIS.

    • Transects every, say, 100 m along a baseline.

  3. For each transect, calculate:

    • Distance between bank positions for different years → bank erosion/accretion.

Output:

  • Erosion map (where bank moved outward)

  • Accretion map (where land gained).

b) Centerline shift

  1. Use the centerlines of different years.

  2. Measure the shortest distance from each vertex of older centerline to newer centerline.

  3. Summarize:

    • Average migration rate (m/year)

    • Max migration at specific bends.

c) Channel width and planform changes

From river polygons for each year:

  • Calculate channel area and mean width (area / centerline length).

  • Derive sinuosity:

    • Sinuosity = channel length / valley length

    • Compare sinuosity between years to see straightening or meandering.

6. Create GIS-Based Change Maps

You can visualize migration nicely:

  1. Overlay multiple year polygons:

    • Use different colors (1990 blue, 2000 cyan, 2010 green, 2020 red).

  2. Use Symmetrical Difference / Union (vector overlay):

    • Identify areas where the river was present in earlier year but not later → abandoned channel / bar formation.

    • Areas where river is new in later year → recent erosion zones.

  3. Make raster of frequency:

    • Convert water polygons of each year to rasters (1 = water, 0 = land).

    • Sum them → cell values show how many years the cell was water.

    • High values → stable channel core, low values → highly mobile margins.

7. Quantify Migration Rate

For a study report or paper, compute:

  • Linear migration rate:

    • Distance moved (m) / time interval (years).

  • Separate rates for:

    • Outer bends vs inner bends.

    • Upstream, midstream, downstream reaches.

Tabulate results, e.g.:

Reach Period Avg. bank shift (m) Rate (m/yr) Dominant process
R1 2000–2010 50 5 L-bank erosion
R1 2010–2020 120 12 R-bank erosion

8. Link Migration to Controlling Factors

With GIS, overlay other layers:

  • Land use / land cover change maps → deforestation, agriculture, sand mining near banks.

  • Infrastructure → roads, embankments, spurs, dams.

  • Slope / geology (from DEM and geological maps).

  • Rainfall / discharge stations (if you have them).

Then you can analyze:

  • Where high migration coincides with:

    • Steep outer banks

    • Certain soil types

    • Human activities (sand mining, removal of vegetation, etc.)

9. Tools / Software You Can Use

  • QGIS (free):

    • Raster calculator → NDWI/MNDWI.

    • Vector editing → digitize banks & centerlines.

    • Plugins:

      • “Profile Tool”, “Semi-Automatic Classification Plugin” for classification.

  • ArcGIS Pro / ArcMap:

    • Spatial Analyst, Image Analyst.

    • DSAS (Digital Shoreline Analysis System) – widely used for shoreline and river bank migration.

  • R / Python (optional, for automation):

    • Libraries: raster, sf, terra in R; rasterio, geopandas, shapely in Python.

    • Can automate NDWI calculation, classification, and distance statistics.

10. Simple Workflow Summary

If you want a quick “step list”:

  1. Collect multi-date satellite images (e.g., 1990, 2000, 2010, 2020).

  2. Preprocess (projection, clipping, georeferencing).

  3. Extract river water polygons using NDWI or manual digitizing.

  4. Generate centerlines and bank lines for each year.

  5. Create transects and measure bank/centerline shift between years.

  6. Calculate migration rate (m/year) and map erosion/accretion zones.

  7. Overlay with land use, infrastructure, DEM to interpret causes.

  8. Prepare maps, graphs, and tables for reporting


ArcGIS gives you a lot of power for this. I’ll walk you through a practical, button-by-button workflow to track river channel migration using ArcGIS (ArcMap/ArcGIS Pro), including DSAS.

I’ll assume:

  • You have multi-date satellite images (e.g., 1990, 2000, 2010, 2020)

  • You want to map and measure how much the river has shifted.

1. Prepare Your Data

1.1 Set up a geodatabase

  1. Open ArcGIS Pro (or ArcMap).

  2. Create a File Geodatabase:

    • Catalog → right-click folder → New → File Geodatabase.

  3. Create feature classes like:

    • River_1990, River_2000, River_2010, River_2020

    • All in the same projected coordinate system (e.g., UTM).

1.2 Import and align imagery

For each satellite image (e.g., Landsat, Sentinel-2):

  1. Add to map: Map → Add Data.

  2. Check coordinate system:

    • Right-click layer → Properties → Source/Spatial Reference.

  3. If needed, Project Raster:

    • Geoprocessing → Project Raster (set same projection as your geodatabase).

  4. Clip image to study area:

    • Data Management → Raster → Raster Processing → Clip.

If you have old maps / aerial photos:

  • Use Georeferencing toolbar to tie to ground control points and save the georeferenced raster.

2. Extract the River Channel for Each Date

You can either digitize manually or use NDWI/MNDWI.

2.1 Option A: Manual digitizing (high control)

  1. Create polygon feature class:

    • In Geodatabase: New → Feature Class → Polygon, name it e.g. River_2000.

  2. Add to map, start editing:

    • Edit tab → Create → Create Features (ArcGIS Pro)

    • Or in ArcMap: Editor → Start Editing.

  3. On-screen digitize water boundary following the river edges for that date.

  4. Save edits and stop editing.

  5. Repeat for each year (1990, 2000, 2010, 2020).

2.2 Option B: Semi-automatic using NDWI

For Landsat (example: Green = Band 3, NIR = Band 5 for Landsat 8):

  1. Go to Geoprocessing → Raster Calculator.

  2. NDWI formula (example for Landsat 8):

    NDWI = ( "Green_band" - "NIR_band" ) / ( "Green_band" + "NIR_band" )

  3. Output a new raster: NDWI_2000.

  4. Classify water:

    • Use Raster Calculator again with threshold, e.g.:

      Con("NDWI_2000" > 0, 1, 0)

    • Name it Water_2000.

  5. Convert water raster to polygon:

    • Conversion → From Raster → Raster to Polygon.

    • Input: Water_2000, field: Value.

    • Only keep polygons where Value = 1 (water).

  6. Clean polygons:

    • Eliminate or Eliminate Polygon Part to remove small specks.

    • Optionally Smooth Polygon (Cartography Tools) for nicer boundaries.

Repeat for each year: Water_1990, Water_2000, Water_2010…

3. Derive Centerlines and Banklines

You’ll use the river polygons to get lines.

3.1 Get banklines

  1. Use Feature To Line:

    • Data Management → Features → Feature To Line.

    • Input: river polygon (e.g., River_2000).

    • Output: Banklines_2000.

  2. Split into left and right bank if needed:

    • Manually split where necessary using Split tool in Edit.

    • Or you can just keep a single bankline per year if you only care about main bank movement.

3.2 Get centerline (optional but useful)

Approach: approximate centerline by manually digitizing along channel midline:

  1. Create line feature class: Centerline_2000.

  2. Start editing and trace the midline along the channel.

  3. Save, repeat for other years.

(You can use advanced tools / skeletonization extensions, but manual works fine for moderate-length rivers.)

4. Measure Migration with DSAS (Digital Shoreline Analysis System)

DSAS is an ArcGIS extension (from USGS) widely used for shoreline and riverbank migration.

4.1 Install and set up DSAS

  1. Download DSAS (if not already): you get a toolbox/add-in.

  2. Add the DSAS toolbar or toolbox in ArcGIS.

  3. Prepare data structure:

    • All banklines for different years in one feature class:

      • Merge → Banklines_AllYears.

    • Add attribute fields like:

      • Date_ (e.g., 2000, 2010, 2020)

      • Source (imagery type).

    • Make sure field names follow DSAS requirements (e.g., an ID field, date field as DATE_ or as specified in DSAS docs).

4.2 Create the baseline

DSAS needs:

  • Baseline: a line roughly parallel to the banklines but landward.

Steps:

  1. Create a new line feature class: Baseline.

  2. Digitize a smooth line along the general direction of the river, slightly inland (for the bank you want to analyze).

  3. Ensure baseline has a unique ID field (e.g., BASE_ID).

4.3 Generate transects with DSAS

  1. Open DSAS Transect Generation tool.

  2. Input:

    • Baseline layer = your Baseline.

    • Shoreline layer = Banklines_AllYears.

    • Transect spacing = e.g., 100 m (distance between profiles).

    • Transect length = enough to cross all bank positions (e.g., 1000 m).

  3. Run the tool:

    • Output: Transects layer.

Each transect will intersect all banklines of different years.

4.4 Calculate statistics (migration rates)

  1. Run DSAS – Rate Calculation tool.

  2. Input:

    • Transects

    • Banklines_AllYears

    • Date field (e.g., Date_).

  3. Choose statistics to compute:

    • SCE (Shoreline Change Envelope) – total movement.

    • EPR (End Point Rate) – simple rate between first and last year (m/year).

    • LRR (Linear Regression Rate) – trend-based rate.

  4. Output: Transects_with_Rates.

Each transect now has attributes like EPR, LRR you can map.

5. Visualize Erosion and Accretion

5.1 Map bankline evolution

  1. Symbolize each year’s bankline with a different color:

    • 2000 = blue

    • 2010 = green

    • 2020 = red

  2. Overlay them to see where the bank moved outward or inward.

5.2 Create change polygons (erosion/accretion)

  1. Use Union or Symmetrical Difference:

    • Analysis → Overlay → Union

    • Input: e.g., River_2000 and River_2020.

  2. The output polygons will show:

    • Areas where river existed before but not after → accreted land (channel abandonment).

    • Areas where river is new in later year → erosion sites.

  3. Add a new field ChangeType and classify polygons manually:

    • Erosion

    • Accretion

    • No Change

6. Calculate Migration Distance and Rate Manually (Without DSAS)

If you don’t want DSAS, you can still do it.

6.1 Create transects manually

  1. Create line feature class Transects.

  2. Draw lines roughly perpendicular to flow at regular intervals.

  3. Use Generate Points Along Lines (if you want sample points).

6.2 Measure distances

  1. Use Near or Generate Near Table:

    • Analysis → Proximity → Near.

    • Input: vertices from older bank/centerline.

    • Near features: newer bank/centerline.

  2. This gives a distance field = migration distance between two years.

Then compute rate:

Rate (m/yr) = Distance (m) / Time interval (years)

You can do this with Field Calculator.

7. Summarize and Export Results

7.1 Summary stats

Use Statistics or Summary in ArcGIS:

  • Average migration rate per reach.

  • Max/min rates.

  • Number of transects in different rate classes (0–5 m/yr, 5–10 m/yr, etc.).

7.2 Export maps/tables

  • Design a layout in Layout View (ArcMap) or Layout (Pro).

  • Export to PDF or PNG.

  • Export tables as .csv for use in Excel or R.


1. Choose Years & Download Data (West Rapti Focus)

a) Decide time steps

For West Rapti, good combos (depending on your story):

  • 1990 / 2000 / 2010 / 2020 (decadal)

  • Or pre-major flood vs post-flood years (e.g., big events you know about from DWIDP/DoI, local records).

b) Get satellite images

Use:

  • USGS EarthExplorer or Copernicus Open Access for:

    • Landsat 5/7/8 (30 m, from 1980s–present)

    • Sentinel-2 (10 m, from ~2015 onwards)

Tips for West Rapti:

  • Pick dry season (Nov–Apr) images when:

    • Cloud is minimal

    • Water level relatively low but channel clearly visible.

For each chosen year:

  • Download 1–2 cloud-free scenes that fully cover your study reach.

If you plan publication: note Path/Row (Landsat), acquisition date, sensor.

2. ArcGIS: Set Up Project for West Rapti

a) Coordinate system

Use a UTM zone covering West Rapti, for example:

  • Project everything to UTM (WGS 84, correct zone for mid–western Nepal).

In ArcGIS:

  1. Create File Geodatabase → WestRapti_Migration.gdb.

  2. Inside, create feature classes:

    • River_1990, River_2000, River_2010, River_2020 (Polygon)

    • Later: Banklines_AllYears, Baseline, Transects, etc.

b) Preprocess imagery

For each year:

  1. Project Raster if needed

    • Geoprocessing → Project Raster → output in UTM.

  2. Clip to your study reach (to make processing faster):

    • Data Management → Raster → Raster Processing → Clip

    • Use a polygon of your AOI buffer along West Rapti.

3. Extract West Rapti Channel (Water) Each Year

Option 1 (Recommended): NDWI method in ArcGIS

Let’s take Landsat 8 example (for ~2013 onwards):

  • Green band = Band 3

  • NIR band = Band 5

3.1 Compute NDWI

  1. Open Raster Calculator.

  2. Expression (Landsat 8):

("LC08_B3" - "LC08_B5") / ("LC08_B3" + "LC08_B5")

  • Output name: NDWI_2015.

(For Landsat 5/7, use Green = Band 2, NIR = Band 4; same formula.)

3.2 Extract water from NDWI

  1. Again Raster Calculator:

Con("NDWI_2015" > 0, 1, 0)

  • Tweak threshold (like >0, >0.1, etc.) by trial and error until it nicely captures the river and avoids fields/sandbars.

  • Output: Water_2015.

  1. Convert to polygon:

    • Conversion → From Raster → Raster to Polygon

    • Input raster: Water_2015, field: Value.

    • Output: WaterPoly_2015.

  2. In WaterPoly_2015 attribute table:

    • Select Value = 1 → these are water polygons.

    • Export Selected as River_2015 (polygon in your geodatabase).

    • Delete small patches with Eliminate or select polygons < certain area and delete.

Repeat same steps for other years: River_1990, River_2000, River_2010, River_2020.

If NDWI is messy in braided sections of West Rapti (e.g., near Rapti–Deukhuri), you can manually clean polygons by editing.

Option 2: Manual digitizing (if imagery unclear)

For each year, open the respective image + Google/Bing basemap, and:

  • Create polygon FC River_Year.

  • Start editing → trace water boundary along West Rapti accurately.

  • Takes time, but good for smaller reaches.

4. Get Banklines & Centerlines for Each Date

4.1 Banklines (left + right)

For each River_Year polygon:

  1. Use Feature To Line:

    • Data Management → Features → Feature To Line

    • Input: River_2015, Output: Banklines_2015.

  2. Add fields: Year (short integer or text), fill 2015.

  3. Optionally, split into Left Bank / Right Bank if you want bank–specific rates:

    • In edit mode, use Split tool and manually separate left/right segments.

    • Add field Bank with values like "L" or "R".

Combine all banklines:

  • Use Merge:

    • Target: Banklines_1990, Banklines_2000, Banklines_2010, Banklines_2020

    • Output: Banklines_AllYears.

Now Banklines_AllYears has:

  • One line per year (or per bank per year)

  • Fields: Year, Bank, etc.

4.2 Centerline (optional but nice for planform)

  1. Create line FC Centerline_Year for each River_Year.

  2. Trace mid-channel by hand along West Rapti.

  3. Later you can measure centerline shift instead of bankline if you prefer.

5. Use DSAS for Quantitative Migration on West Rapti

This is powerful and looks professional in a thesis.

5.1 Prepare DSAS shoreline layer

DSAS expects:

  • One shoreline layer with all dates + a date field.

We already have Banklines_AllYears. Make sure:

  • It has a unique ID (e.g., ID field).

  • A date field (e.g., Date_):

    • Store actual dates like 1990-01-01, 2000-01-01, etc.,

    • or at least numeric year (then follow DSAS format guide).

5.2 Create Baseline

Baseline = line roughly parallel to river but landward.

  1. Create line FC Baseline.

  2. Decide which bank you’re analyzing (e.g., left bank along Deukhuri side).

  3. Digitize baseline along that side:

    • Keep longer and smoother than individual bends.

    • Baseline should not cross the banklines.

5.3 Generate Transects in DSAS

In DSAS toolbox:

  1. Generate Transects:

    • Baseline layer = Baseline

    • Shoreline layer = Banklines_AllYears

    • Transect spacing = e.g., 100 m or even 50 m if reach is not too big.

    • Transect length = enough to cross all years’ banklines (e.g., 1500 m).

  2. Run → you get Transects feature class:

    • Each transect has a unique ID and intersects all banklines.

5.4 Compute Migration Rates

Use DSAS Calculate Rates tool:

  1. Input transects = Transects.

  2. Shoreline layer = Banklines_AllYears.

  3. Date field = your Date_ field.

  4. Choose methods:

    • EPR (End Point Rate) = (distance between first and last shoreline) / time.

    • LRR (Linear Regression Rate) for more years.

    • Optional: SCE (max change width).

Output: Transects_Rates with fields:

  • EPR (m/yr)

  • LRR (m/yr)

  • SCE (m)

  • etc.

Now you can map where West Rapti is migrating fastest.

6. Visualize Erosion / Accretion along West Rapti

6.1 Rate map (from DSAS)

  1. Symbolize Transects_Rates:

    • Graduated colors based on EPR:

      • < 0 m/yr = accretion (bank prograding)

      • 0 m/yr = erosion (bank retreat).

  2. Break into classes:

    • 0–2 m/yr (low)

    • 2–5 m/yr (moderate)

    • 5 m/yr (high)

This gives a beautiful along-river bar chart map of migration.

6.2 Erosion vs new land polygons

Between two specific years, say 2000 and 2020:

  1. Take River_2000 and River_2020.

  2. Use Union:

    • Analysis → Overlay → Union

    • Output: Union_2000_2020.

  3. In attribute table, define ChangeType:

    • If polygon area was water in 2000 but not in 2020 → Abandoned channel / Accretion land.

    • If polygon area was not water in 2000 but water in 2020 → New erosion zone.

  4. Symbolize:

    • Erosion = red

    • Accretion = green

    • Stable = transparent/light.

This works nicely near meander bends of West Rapti to show where it has eaten farmland or created new bars.

7. Add Context Layers (for Discussion in Thesis)

In ArcGIS, overlay:

  • Land use / land cover (e.g., forest, agriculture, settlement)

  • Roads, embankments, spurs (from OpenStreetMap or DWIDP)

  • DEM/contour (SRTM/ALOS) to show valley shape and floodplain width.

Then you can analyze:

  • Higher migration where:

    • Bank is unconsolidated alluvium

    • Sand mining or agriculture extend up to bank

    • No protective embankments.

This will strengthen your interpretation part (“Why is West Rapti migrating more in this reach?”).

8. Suggested Outputs for Your West Rapti Study

For a report/article:

  1. Map 1 – Location map of study reach on West Rapti.

  2. Map 2 – Multi-year river positions (e.g., 1990, 2000, 2010, 2020) overlaid.

  3. Map 3 – Erosion/accretion polygons (e.g., 2000–2020).

  4. Map 4 – Transect-based migration rate (EPR or LRR) along river.

  5. Graph – Histogram of migration rate (m/yr).

  6. Table – Average migration rate by reach (upper, middle, lower West Rapti).

If you want, next I can:

  • Help you write a methodology section specifically titled
    “River Channel Migration Analysis of West Rapti River using GIS and Remote Sensing”
    with proper steps and technical terms suitable for a journal paper (like Discover Water / Springer),

  • Or help you design figures (map layouts + caption style) and equations to include.

Comments

Post a Comment