Introduction

The advent and evolution of ancient human civilization and modern settlements occurred on the banks of small streams and tributaries in the large river basins1,2,3. Moreover, such small natural streams also delineate unique global biodiversity hotspots4,5,6,7. However, the unprecedented population growth8 is responsible for land use and land cover (LULC)9 change, which disproportionately and irreversibly diverted the land under surface water bodies into agriculture, settlement, infrastructure development, and industries10. Additionally, the recent climate scenarios11, including droughts and floods in the transboundary river basins12, intensive farming in the riverbeds and wetlands of Europe and Southeast Asia13,14,15, and the intensification of climate extremes16,17 further irreversibly altered the natural stream networks. As a result, the small streams in the large river basins are disappearing, and there is no update on their historical and current morphometric status. Hence, field inventory and visual inspections of morphometry have become inevitable in understanding the shrinking and disappearing trends in order, number, and length of small and large stream(s) in small watersheds.

The morphometric parameters define a basin’s topographical, geological, and hydrological condition18,19. The digital elevation model (DEM) is a widespread technique in the terrain visualization, processing, and quantification of topographic attributes in morphometric analyses to understand river basin structure and functions at the local20,21,22,23, regional24,25,26, and global scales27.

Despite spatially explicit morphometric analysis28, the DEM cannot delineate and portray the actual status of the watersheds and subwatersheds corresponding to changes in land use and land cover (LULC), including decreasing the number and length of streams. DEM-based morphometric studies have limitations due to the inconsistent DEM resolutions across global studies and varying data acquisition and processing techniques (i.e., stream thresholds), complicating data comparison and reducing accuracy29. Additionally, DEM-derived simulations often require noise removal in slope data, and the grid size significantly impacts the representation of land surfaces and topographic attributes30,31. High-quality DEMs provide detailed conditioning factors that may not always yield precise information for multivariate modeling, especially when the resolution and size of landforms are critical to the analysis32,33. However, tracing Survey of India (SOI) toposheets with field verification is highly accurate but labour-intensive, especially for large basins like the Ganga34. To overcome these major limitations of the DEM35,36 it is essential to develop and adopt robust computational and field techniques and delineate the historical or reference and current year morphometric status of the small watershed(s) of varying stream networks.

In our study, we addressed the rejuvenation of large river basins by delineating reference year (1970–1977) and current year (2021–2024) morphometry of small watersheds and fifty-six subwatersheds using SOI toposheets, remote sensing, and drone imagery, enabling efficient and detailed ground truthing. The morphometric comparisons (1) explain the mechanism of disappearing and shrinking stream channels of the seven watersheds and fifty-six subwatersheds, (2) help define their threatened category, and (3) delineate ex-situ and in-situ restoration and conservation techniques while considering competing and conflicting land uses14,36. The seven small watersheds were selected based on closeness to (1) ancient rural civilization and supporting traditional livelihood and local biodiversity, (2) ancient cultural civilization and declared as World Heritage Sites (WHS), and (3) ancient industrial sites and supporting provisional and regulatory services in the Ganga River basin (GRB).

Materials and methods

Study area

The location and current drainage order maps of seven small rivers in the Ganga River basin are given in Fig. 1. These seven rivers were selected based on the following criteria: (i) closeness to ancient rural civilization and supporting traditional livelihood and local biodiversity, (ii) closeness to ancient cultural civilization and declared as World Heritage Sites (WHS), and (iii) closeness to ancient industrial sites and supporting provisional and regulatory services (Supplementary Table S1). Additionally, on the basis of dominant livelihood practices, seven watersheds were classified as (1) agricultural watersheds (Bodla, Banki, and Danro), (2) cultural watersheds (Khudar and Urmil), and (3) industrial watersheds (Katri and Khudia). The Bodla watershed is historically one of the oldest 4th -order running water systems that witnessed the dawn of human civilization in the upstream Betwa River basin, central India. The Khudar and Urmil watersheds (tributaries of Ken River) have seen the evolution of WHS Khajuraho, the ancient city of medieval India, in the marginal plain of the Ganga River basin. The Banki and Danro are tributaries of the North Koel River and distinctly unique watersheds linking the Gondwana with the Vindhyan Supergroup. The catchments of the Katri and Khudia Rivers are highly populated by ancient coal mining (Jharia and Raniganj Coalfields) and allied industries in the Barakar formation of the Gondwana system.

Computational techniques

The downloaded reference year (1970–1977) SOI toposheets37 (Supplementary Table S1) were georeferenced based on Universal Transverse Mercator (UTM) projection and the World Geodetic System (WGS) and mosaicked using Arc GIS 10.2 software. The mosaic toposheets were manually digitized for the drainage network extraction using the Arc GIS 10.2 editing tool. Further, the digitized shapefiles were converted into a topology for the error correction38 and then filled attributes for each stream based on the Strahler method of stream ordering39. The un-branched streams were designated as 1 st order streams; two 1 st order streams joined to form 2nd order streams, the joining of two 2nd order streams resulted in a 3rd order stream, and so on10. The watershed and sub-watershed delineation were carried out with the help of pour points in the Arc Hydro tools using ASTER (Advanced Spaceborne Thermal Emission and Reflection and Radiometer) 30 m spatial resolution digital elevation data set (downloaded from https://search.earthdata.nasa.gov/search)40 and cross-checked with toposheet contours (Supplementary Table S1).

Field inventory

The field inventory updated the current morphometric details using a global positioning system (GPS, Model Garmin eTrex 30) and Drone surveying (DJI AIR 2 S)10 of seven watersheds and fifty-six subwatersheds. The qualitative and quantitative verifications involved the presence/absence of streams along with their origin and confluence points. The recorded field data were used to delete or retain streams in the respective digitized reference shapefiles and depict changes in the drainage network(s). Associated spatial data were generated by repeating the GIS processing exercises using Arc GIS 10.2 software.

Morphometric analysis

The basin length (Lb), area (A), perimeter (P), stream order (1st, 2nd, 3rd, 4th, 5th, 6th, etc.), stream number (Nu), and stream length (Lu)39,41,42,43 was determined (Supplementary Table S2) for reference and current morphometry of seven watersheds and fifty-six subwatersheds. These primary morphometric parameters were used to analyze the linear aspects (total stream length, mean stream length, bifurcation ratio, and mean bifurcation ratio)43 (Supplementary Table S3) and drainage texture (drainage density, drainage density ratio, drainage frequency, drainage texture, constant of channel maintenance, and infiltration number)10,41,42,44,45 (Supplementary Table S2 and S4). Statistical analysis of data was performed by using IBM SPSS version 22.0 software46.

RRLCC analysis

We used the “River Red List Categories and Criteria (RRLCC),” which deals with the percent change in the Nu, Lu, and Dd of the watersheds and sub-watersheds10 (Table 1 and Supplementary Table S5). The RRLCC classified seven watersheds and their fifty-six subwatersheds into the following threatened categories: not evaluated (NE), data deficient (DD), least concern (LC), near threatened (NT), vulnerable (VU), endangered (EN), critically endangered (CR), near to extinction (NEX), and extinct (EX)10 (Supplementary Table S6).

The decision to initiate ex-situ conservation (agricultural land, village areas, peri-urban settlement, urban municipalities, industrial hubs, etc.) for the streams in a watershed and its subwatersheds should be based on the following criteria: (1) when a certain watershed is vulnerable to significant anthropogenic activities or stochastic events that threaten its ecological, geological, and geomorphological integrity, (b) when the watershed or its subwatersheds have reached a critically endangered (CR) status, near to extinction (NEX), or have become extinct (EX) within a relatively short period. Additionally, special consideration for ex-situ conservation is to be taken in case of threat to the cultural, economic, or scientific importance of the region47. Similarly, the in-situ conservation (protected areas like wildlife sanctuaries, national parks, biosphere reserves, world heritage sites, etc.) criteria for the streams of a watershed should involve a comprehensive assessment of habitat protection, stream viability, threat analysis, legal frameworks, and continuous morphometric monitoring to preserve and maintain the stream diversity and ecological functionality of the watershed. These attributes are defined to support the ex-situ and in-situ conservation of watersheds, which also includes the conservation of wild species in their habitat48,49, sustainable utilization of biodiversity resources49, and equitable sharing of benefits49.

LULC analysis

The unsupervised classification of the Landsat 5 and Landsat 8 (spatial resolution of 30 m) satellite datasets using an iterative self-organizing data analysis technique algorithm (ISODATA)50 was adopted to generate land use and land cover (LULC) change scenarios in the seven watersheds10. In our study, the land features were classified into five LULC classes: (a) agriculture (lowland and highland crop fields with and without crops); (b) barren (dry with a few scattered herbaceous plants and no trees); (c) built-up (high, medium, and low-density settlements, dispersed settlements, infrastructures such as schools, hospitals, industries, bridges, and roads); (d) vegetation (forest cover, trees outside forests, road plantation, shrubs, and herbaceous layer); and (e) water (wet and dry rivers, river banks, waterlogged areas, and small ponds)10,51. The accuracy assessment of the LULC classifications was done by the user accuracy (UA), producer accuracy (PA), overall accuracy (OA), and kappa coefficient52,53 (Supplementary Table S2). The validation data for the reference year (1991–1995) was obtained from high-resolution satellite imagery and for the current year (2020–2024) was obtained from the ground truth points.

Results and discussion

Morphometric comparison of linear aspects

The results of linear morphometric parameters are given in Table 1 and Supplementary Table S3. The seven watersheds and their fifty-six subwatersheds showed an extreme loss in the number of streams (Nu) and the total length of streams (Lu). The comparison of reference and current morphometry dominated by agriculture activities (Bodla, Banki, and Danro) showed ~ 83–89% and ~ 53–70% loss in the Nu and Lu, respectively; coal mining dominated small rivers (Katri and Khudia) showed ~ 60–75% and 33–55% loss in the Nu and Lu, respectively; and the watersheds characterized by the establishment of cultural sites (Khudar and Urmil) showed ~ 46–64% and ~ 40–54% loss in the Nu and Lu, respectively.

The 1 st, 2nd, and 3rd order streams showed maximum loss in their Nu and Lu across upstream, midstream, and downstream regions of the watersheds and subwatersheds (Fig. 2, Supplementary Fig. S1). The higher decrease in the agricultural watersheds than in the cultural and industrial watersheds indicates recent LULC changes in response to implementing India’s national land utilization policy (2013)54. The LULC changes showed 16.82–82.46% and 9.79–390.40% increase in the agriculture and built-up areas, respectively, while 20.75–76.82%, 15.97–31.40%, and 43.79–69.29% loss of watershed areas under barren measures, vegetation, and water bodies, respectively (Fig. 3, Supplementary Table S7 and S8). These geomorphological losses converted 3rd, 4th, 5th, or 6th-order rivers(s) to 2nd, 3rd, 4th or 5th-order rivers(s). Stream order shifting in subwatersheds is the hydrogeomorphological evidence of decreasing stream diversity (Supplementary Fig. S1). The increase in the mean bifurcation ratio (Rbm) and decrease in the coefficient (ρ) showed the prevalence of anthropogenic alterations of geologic structures and geomorphologies, especially their agricultural diversions (Supplementary Fig. S2), resulting in shallowing and shrinking storage capacity of the watersheds and subwatersheds10. Such alterations in the primary linear parameters reveal a disproportionately irreversible impact on the linear aspects of the subsequent high-order tributaries (e.g., Betwa, Ken, North Koel, Barakar, and Damodar) in the Ganga River basin10.

Morphometric comparison of drainage texture

The results of the drainage texture analysis are given in Supplementary Table S4. The spatial variability in the Dd is significant within and between seven watersheds (P < 0.05) (Supplementary Table S9). The mean Dd of seven watersheds and fifty-six watersheds decreased from 2.30 km/km2 to 0.90 km/km2 in the last 50 years. The seven watersheds showed a 33.33–70.16% decrease in the drainage density (Dd) (Table 1), indicating spacing between streams of different orders increased from the reference year (1970–1977) to the respective current year (2021–2024). The agricultural watersheds (Bodla, Banki, and Danro) showed a higher loss in the Dd (> 50%). Additionally, the recent LULC changes in the Khudia (diversion of forest area for rainfed agriculture and coal mining and industrial purposes) and Urmil watersheds (stone-sand-soil mining, brick industries, reservoir construction, irrigated farming, settlements, etc.) led to > 50% decrease in the Dd. The Katri and Khudar watersheds showed < 50% loss in the Dd, indicating both have suffered massive LULC changes due to coal mining and cultural civilization, respectively, in the past. The decreasing Dd indicates the diversion of undulating hilly pediplain into agricultural land, settlement, and infrastructure development (Supplementary Fig. S2) and increases the subsoil permeability with low surface runoff to the main trunk10. In the seven watersheds and their subwatersheds, DdR values ranged from 1.08 to 5.64. A drainage density ratio (DdR) close to 1 represents an ideal equilibrium watershed condition9. The DdR of > 1 reveals decreasing stream order connectivity and reduced hydrological flow and is a valuable indicator of morphometric alterations in the changed LULC scenario. The significant decrease in the stream frequency (Fs), drainage texture (Dt), and infiltration number (If) correspond to disappearing dendritic morphometry, porous surface, and low runoff conditions, respectively, causing reduced river flow. These results in the high constant of channel maintenance (C), corresponding to the high infiltration capacity of the altered surface and subsurface profiles.

Assessment of threatened categories

The results of the “River Red List Categories and Criteria (RRLCC)” of seven watersheds are given in Table 1. The mean of three criteria (Nu, Lu, and Dd) varied from 63 to 77%, indicating two agriculture watersheds (Bodla and Danro) are “Endangered (EN)” while the third watershed (Banki) is “Critically Endangered (CR)”10. The 43.2 to 61.8% and 41.8 to 57.2% variations in the mean of three criteria classified industrial (Katri and Khudia) and cultural watersheds (Khudar and Urmil) into “Vulnerable (VU) and Endangered (EN)” categories. Similarly, 5, 6, 21, 17, and 7 subwatersheds were found in the “Near Threatened (NT), Vulnerable (VU), Endangered (EN), Critically Endangered (CR), and Extinct (EX) categories,” respectively (Supplementary Table S5). These RRLCC categories substantiate the alarming geomorphological shifting towards the extinction of stream morphologies in watersheds that need regulatory attention at par with the conservation of the “IUCN Red Data List of Threatened Species’’55. Before this study, seven watersheds and fifty-six subwatersheds were “Not Evaluated (NE), Data Deficient (DD), and Least Concern (LC).” However, after detailed ground truth verification of the existing morphometry, they are data-rich and highly concerned at the microscale level in the Ganga River basin because ~ 68% of subwatersheds are in endangered and critically endangered categories. The colour code for the river red list (RRL) is adapted from the IUCN Red List55 except for LC which is given the blue colour code to represent the water body. The RRLCC is central dogma to ex-situ (agricultural land, village areas, peri-urban settlement, urban municipalities, industrial hubs, etc.) and in-situ (protected areas like wildlife sanctuaries, national parks, biosphere reserves, world heritage sites, etc.) conservation of small rivers in the transboundary river basins. Additionally, this study developed the latest maps of the seven small watersheds and their fifty-six subwatersheds to exemplify the IUCN-protected areas (PA), including fresh waters48.

The RRLCC comparison of seven watersheds and fifty-six subwatersheds reveals that ancient civilization and livelihood practices (including settlement, mining, agriculture, trade, etc.) were sustainable, as evident from high Nu, Lu, and Dd extracted from their reference toposheets. The recent LULC changes in the last 50 years to meet the water-energy-food demands of the growing population resulted in an unprecedented loss in the morphometric attributes (Nu, Lu, and Dd). For example, 843 BCM water supply for 4,17,668 MW energy and 329,687 MT food production56,57,58 resulted in 1545 m3 per capita water availability in water-stressed India59,60. At present, the dynamics of groundwater resources in the seven watersheds are varied from semi-critical (Bodla, Banki, Danro, Khudar, and Urmil) to overexploited (Katri and Khudia)61.

The National Forest Policy of India (1952) adopted the allocation of 33% of the geographical area under forest cover, including balanced and complementary land use62. Similarly, the National Water Policy (2012)36 needs to be revised, and commitments must be made to allocate statutory land to restore and conserve small streams or rivers irrespective of their length, width, and depth along their source to mouth in large watersheds10. Nevertheless, a ray of hope was reflected when the UN Water Conference in New York (2023) announced the freshwater challenge aims to restore 300,000 km of degraded rivers by 203063. Given this, the 9,493 total reference Nu and 7,006 km total reference Lu in the 3433.85 km2 area of seven small watersheds decreased to 2,473 streams and 3,084.71 km, respectively, in the last 50 years. Consequently, the Dd showed a 56% loss from 2.04 km/km2 to 0.90 km/km2. The reference Nu and Lu are ~ 2.75 and ~ 2.0 times greater than the area of seven watersheds (3433.85km2) (Table 2); hence, the reference Nu and Lu in the 8,61,404 km2 of the Ganga River basin57 could be ~ 23,68,861 streams and ~ 17,57,264 km, respectively (Table 2). From this empirical estimate, it is required to rejuvenate ~ 56% of lost streams of different orders (9,84,068 km) in the Ganga River basin to achieve restoration of drainage density ≥ 2 km/km2. Moreover, the UN’s commitment to restoring 300,000 km of degraded rivers needs 150,000 km2 of watershed areas for the ex-situ and in-situ restoration of three intrinsic morphometric attributes: stream number (Nu), stream length (Lu), and drainage density (Dd). This will help link the restoration and conservation of disappearing and depleted water resources with other embedded tipping elements in the Earth’s climate46,64,65,66,67,68,69,70,71. In a nutshell, ex-situ and in-situ conservation of streams of small watersheds and subwatersheds at the local levels can minimize transboundary conflicts and help equitable multi-stakeholder sharing of water resources.

Table 1 River red list category and criteria (RRLCC) of the seven watersheds: (i) percentage change in morphometric parameters (Nu, lu, and Dd) with respect to reference year and (ii) categorising seven watersheds into different RRLCC categories10 and colour codes (except LC which is given the blue colour code to represent the water body) at par with the conservation category of the “IUCN red data list of threatened Species55 in the reference (1970–1977) and current (2021–2024) assessment years.
Full size table
Table 2 Comparison of (a) summed morphometric attributes of seven watersheds. Based on these empirical estimates the nu, lu, and Dd determined for the (b) GRB and (c) un’s commitment to restoring 300,000 km of degraded rivers63. RY = reference year (1970–1977) and cy = current year (2021–2024).
Full size table

Conclusion

The present study explicitly deciphered the mechanism of small river extinctions in large river basins. The recent gradual and random LULC changes are responsible for the loss of Nu, Lu, and Dd and highlight the unscientific utilitarian approach to exploiting water resources in agricultural, cultural, or industrial watersheds. These irreversible alterations lead to the extinction of small streams and the shrinking of the large river basins. As a result, seven small watersheds and their fifty-six subwatersheds are classified as “Near Threatened, Vulnerable, Endangered, Critically Endangered, or Extinct.” The field-based inventory of watershed and subwatershed morphometry revealed the limitations of DEM-based morphometric delineation and validation. The decreasing Nu, Lu, and Dd of the seven watersheds and fifty-six subwatersheds are surrogate measures of the decreasing Dd of the whole Ganga River basin. The reference and current morphometry data can be revisited by 2030 to understand the annual and decadal changes in drainage density. Therefore, adopting sustainable methods, including a statutory land allocation for the ex-situ and in-situ conservation, is the most required amendment and commitment in the national and international water policies to restore and conserve natural boundaries of streams in small watersheds and their long history of providing ecosystem services in the large river basins.

Fig. 1
figure 1

The location maps of seven watersheds with current year stream networks (2021–2024) in the Ganga River Basin, India (a-b); the seven watersheds are classified into three categories on the basis of dominant livelihood practices i.e. agricultural: (c) Bodla (2022), (d) Danro (2022), and (e) Banki (2021); cultural: (f) Khudar (2024) and (g) Urmil (2024), and industrial: (h) Khudia (2023) and (i) Katri (2023). The shapefile was obtained from open-source website (https://onlinemaps.surveyofindia.gov.in/Home.aspx) to generate map (a). The maps were created using ArcGIS 10.2 software and the website is https://desktop.arcgis.com/.

Full size image
Fig. 2
figure 2

Comparative drainage order change map between reference year (1970–1977) and current year (2021–2024) of the seven watersheds: (a) Bodla (i-ii), (b) Banki (i-ii), (c) Danro (i-ii), (d) Katri (i-ii), (e) Khudia (i-ii), (f) Khudar (i-ii), and (g) Urmil (i-ii); the maps show decreasing stream order(s) from the reference year to current year due to the prevalence of anthropogenic alterations of geologic structures and geomorphologies, especially the agricultural diversions resulting in loss of stream number (Nu) and stream length (Lu). The maps were created using ArcGIS 10.2 software and the website is https://desktop.arcgis.com/.

Full size image
Fig. 3
figure 3

The land use and land cover (LULC) change analysis between the reference year (1991–1995) and current year (2020–2024) of seven watersheds: (a) Bodla (1995–2020), (b) Khudar (1994–2024), (c) Urmil (1994–2024), (d) Banki (1991–2021), (e) Danro (1992–2022), (f) Katri (1993–2023), and (g) Khudia (1993–2023). The increase in the anthropogenically modified agricultural and built-up areas and decrease in barren land, vegetation, and waterbodies resulting in loss of the stream morphological diversity.

Full size image