论文信息

Felix Bauer / Clara Hoffmann

Aquatic Ecosystems and Environmental Frontiers

论文解读

AI 辅助整理
方法
遥感、GIS、景观指标、最小成本路径与图论,100 km²景观。
发现
58.33%生境斑块孤立,修复情景使连通概率从18.60%升至31.40%。
机会点
可将该连通性框架与农业面源污染、湿地保护结合,研究农业景观格局对水生生物多样性的影响。

论文摘要

池塘、水罐、湿地和小型湖泊是承载水生生物多样性的淡水生态系统实例。然而,农业、城市开发和排水系统持续导致这些生态系统碎片化。本研究基于景观方法开发了一种空间方法学,运用遥感、地理信息系统(GIS)、景观指标、最小成本路径分析和图论,对100 km²景观中孤立淡水生境之间的连通性进行制图。共制图24个斑块,占景观总面积的3.84%,其中58.33%的斑块被认定为孤立。识别出9条最小成本廊道,生境斑块H03和H10被确定为重要廊道。生态系统恢复情景将连通性概率从18.60%提高至31.40%。该框架为廊道恢复和保护优先序的确定提供了支持。

全文

Aquatic Ecosystems and Environmental Frontiers

Aquatic Ecosystems and Environmental Frontiers, Volume 4, Issue 4, December 2026 ISSN: 3049-186X
DOI: https://doi.org/10.70102/AEEF/V4I4/AF260431
www.aquaticfrontiers.com 262
Mapping Ecological Connectivity among Isolated
Freshwater Habitats Using Landscape-Based Spatial
Analysis

1* Felix Bauer, Rhine Centre for Professional Excellence, Cologne, Germany.
2 Clara Hoffmann, Rhine Centre for Professional Excellence, Cologne, Germany.

Abstract: Ponds, tanks, wetlands and small lakes are examples of freshwater ecosystems which house aquatic
diversity. However, agriculture, urban development and drainage systems continue to cause fragmentation of these
ecosystems. In this study, a spatial methodolo gy based on a landscape approach that employs remote sensing, GIS,
landscape metrics, least cost path analysis and graph theory was developed to map connectivity between isolated
freshwater habitats in a 100 km² landscape. A total of 24 patches amounting t o 3.84% of the total landscape were
mapped, while 58.33% were found to be isolated. There were nine least cost corridors and habitat patches H03 and
H10 were identified as important corridors. An ecosystem restoration scenario increased the Probability of
Connectivity from 18.60% to 31.40%. The framework supports corridor restoration and conservation prioritization.
Keywords: Freshwater Habitats; Ecological Connectivity; Landscape Fragmentation; GIS; Remote Sensing;
Least-Cost Path; Habitat Corridors; Spatial Analysis.
(Submitted: September 14, 2026; Revised: October 16, 2026; Accepted: December 04, 2026; Published: December 30, 2026)
I. Introduction
Background of the Study
Freshwater ecosystems like lakes, ponds, wetlands, rivers, and reservoirs are important ecological
components providing diverse aquatic life, maintaining ecological balance, and offering a range of
ecosystem services (Field & Parrott, 2022) . These aquatic ecosystems are involved in water regulation,
nutrient cycling, sediment trapping, and supporting both aquatic and terrestrial food chains (Erős et al.,
2012). Nevertheless, due to expanding agricultural lands, urbanization, construction, changes in drainage,
and landscape transformation, freshwater landscapes are being strongly modified. These anthropogenic
impacts result in dividing formerly contiguous aquatic ecosystems into fragmented spatially isolated habitat
patches.
The concept of ecological connectivity refers to the ability of aquatic organisms and ecological processes
to move between habitat patches across a landscape (Luque et al., 2012; (Matos et al., 2019) . The first
concept of connectivity includes structural connectivity which is connected with the distribution of habitats
within the landscape and their proximity. Functional connectivity deals with specific movements of species
and landscape permeability (Ding et al., 2025) . Fragmentation and spatial isolation can restrict dispersal,
limit genetic exchange, reduce recolonization following local disturbances, and increase the probability of
population decline. These effects are particularly important for small ponds, seasonal wetlands, and isolated
freshwater habitats that depend on connections with surrounding aquatic ecosystems for long -term
ecological persistence (Miao et al., 2019).
Habitat assessment for freshwater habitats traditionally focuses on factors such as habitat size, patch
density, and Euclidean distance without taking into account land cover features, artificial barriers, and
landscape resistance (Johnson & Host, 2010; McCullough et al., 2019). The above methods will thus likely
understate the ecological isolation experienced by some organisms while ignoring movement corridors. The
inclusion of indicators of spatial fragmentation, connectivity analysis, modeling of landscape resistance,
and corridor identification in assessing freshwater habitats will give an in -depth insight of the habitats

Aquatic Ecosystems and Environmental Frontiers, Volume 4, Issue 4, December 2026 ISSN: 3049-186X
DOI: https://doi.org/10.70102/AEEF/V4I4/AF260431
www.aquaticfrontiers.com 263
(Bishop-Taylor et al., 2015) . This integrated framework supports systematic conservation prioritization,
habitat restoration planning, and sustainable freshwater landscape management (Phillipsen et al., 2015).
Research Gap
The current study was triggered by four gaps. First, there is no integration of the distribution of
freshwater habitats and the fragmented landscape. Second, the concept of landscape resistance is
overlooked during assessments of connectivity of isolated w ater bodies. Third, key habitat patches for
restoration are not detected on the basis of connectivity analysis (Bishop-Taylor et al., 2015) . Fourth,
restoration corridors connecting isolated freshwater habitats are not prioritized sufficiently in a spatial
context. Filling these gaps requires an integrated approach based on remote sensing, landscape metrics,
least cost path modeling and graph analysis.
Research Objectives
There were four main aims for this study. The first one was the delineation and mapping of the isolated
freshwater habitats through remote sensing and GIS. The second one was the quantification of landscape
fragmentation through spatial landscape metrics. The third one was the estimation of ecological
connectivity through least cost paths and graphs. The fourth one was the identification of priority freshwater
habitats and ecological corridors for conservation and restoration. These objectives provide an or dered
sequence that guides the presentation of the results.
Research Questions
Four research questions were developed to guide the analysis. What are the spatial patterns of freshwater
habitat distribution in the study landscape? What are the most significant land cover properties for habitat
fragmentation? What are the key freshwate r habitat patches that provide connectivity for the movement?
Where conservation efforts need to focus in order to enhance connections among separated water bodies?
Answering the above questions will create a link between spatial distribution of habitats and the processes
involved in movement. This way one can provide the managers with information on how to focus
conservation efforts.
This paper is further divided into four sections. Section 2 presents the study area, spatial data used in
the study, pre-processing of satellite imagery, landscape fragmentation indices, least cost paths and graph
analysis for habitat connectivity. In section 3, habitat distribution, fragmentation, ecological corridors and
key patches of habitat connectivity will be discussed. Ecological implications, corridors significance and
conservation efforts will be discussed in section 4. Finally, Section 5 concludes the study by summarizing
key findings, limitations, and future research directions.
II. Materials and Methods

Figure 1: Conceptual Framework for Mapping Ecological Connectivity Among Freshwater Habitats
Inputs
Remote Sensing
(Land Cover)
GIS Data
(Habitats,
Resistance)
Analysis
Landscape Metrics
(Fragmentation)
Least-Cost Path
(Corridors)
Graph Analysis
(Key Habitats)
Outputs
Habitat
Connectivity
(Habitats,
Corridors, Key
Patches)
Outcome
Conservation
&
Restoration
Prioritization

Aquatic Ecosystems and Environmental Frontiers, Volume 4, Issue 4, December 2026 ISSN: 3049-186X
DOI: https://doi.org/10.70102/AEEF/V4I4/AF260431
www.aquaticfrontiers.com 264
Figure 1 demonstrates the framework designed for the integration of remote sensing and GIS with the
spatial analysis based on the landscape approach. The landscape metrics provide information on the
fragmentation of the habitat, the least cost path models h elp to identify the potential ecological corridors
and the graph -based analysis helps to determine the critical habitat patches, facilitating informed
conservation planning, habitat restoration, and sustainable freshwater ecosystem management.
Study Area
The experiment was carried out in a freshwater landscape of area 100 sq km located in Tamil Nadu,
India. This landscape contains several isolated ponds, irrigation tanks, small wetlands, and various drainage
structures connecting them. The landscape has a semi-arid to sub -humid tropical climatic condition
receiving rainfall mainly from the northeast monsoon season, with gentle undulating topography suitable
for irrigation by tanks. Agricultural lands form the major portion of the matrix with villages and ro ads, as
well as barren areas and scrub land. Major anthropogenic pressures include crop expansion, village and
road development, groundwater extraction and modification of natural drainage lines, all of which influence
the size, isolation and persistence of freshwater habitats.
Data Collection and Sources
The spatial data utilized is listed in table 1 below. Satellite images from Sentinel -2 were used for
mapping waterbodies and vegetation cover (McCullough et al., 2019), while satellite images from Landsat
9 were used for land cover classification. The SRTM digital elevation model was used for terrain and
drainage modeling, while the road network was derived from OpenStreetMap to act as barriers. The images
were collected during a similar period to ensure uniformity in water extents.
Table 1: Spatial Datasets and their Applications
Dataset Source Resolution Application
Sentinel-2 imagery Copernicus 10 m Water and vegetation mapping
Landsat 9 imagery USGS 30 m Land-cover classification
Digital elevation model SRTM 30 m Elevation and terrain analysis
Land-use data Satellite classification 10–30 m Landscape resistance
Waterbody boundaries GIS digitisation Variable Habitat patch identification
Road network OpenStreetMap Vector Barrier identification
Drainage network DEM/GIS extraction 30 m Hydrological pathway mapping
Satellite Image Preprocessing
The cloud minimized images were then processed with atmospheric corrections and cloud masking
before appropriate spectral bands were chosen. The waters were highlighted through the use of Normalized
Difference Water Index (NDWI) and Modified Normalized Dif ference Water Index (MNDWI) as shown
by equations 1 and 2. Afterwards, land classification and habitat boundary extraction were done, and results
were verified through independent validation samples (Phillipsen et al., 2015; Ding et al., 2025). Thresholds
for water bodies were set locally due to variability in thresholds based on turbidity, vegetation cover, and
season.
NDWI =
Green−NIR
Green+NIR (1)
MNDWI =
Green−SWIR
Green+SWIR (2)
Land-Use and Land-Cover Classification
Landscape was categorized into five classes, namely agricultural land, built-up land, natural vegetation,
fresh water, and barren or open land (Table 2). Supervised classification model was trained using sample
images; accuracy assessment was performed using error matrix (Field & Parrott, 2022). Agricultural land

Aquatic Ecosystems and Environmental Frontiers, Volume 4, Issue 4, December 2026 ISSN: 3049-186X
DOI: https://doi.org/10.70102/AEEF/V4I4/AF260431
www.aquaticfrontiers.com 265
made up 42.10% of the landscape, while fresh water constituted only 3.84% of the total area, which clearly
shows that aquatic habitat is a smaller component of the landscape.
Table 2: Land-Use and Land-Cover Classification of the Study Landscape
Land-cover category Area (km²) Area (%) Ecological relevance
Agricultural land 42.10 42.10 Potential runoff and habitat pressure
Built-up area 18.00 18.00 High artificial disturbance
Natural vegetation 25.00 25.00 Movement-supporting matrix
Freshwater bodies 3.84 3.84 Primary aquatic habitats
Barren/open land 11.06 11.06 Variable movement resistance
Total 100.00 100.00
Landscape Fragmentation Analysis
The fragmentation metrics were based on the following calculations for patches number (NP), patches
density (PD), mean patch area (AREAMN), edge density (ED) and largest patch index (LPI), provided in
table 3 (Miao et al., 2019). The following formulas include: N – patches number, A – the area of the whole
landscape, AT – the area of total freshwater habitat, Amax – the area of the largest patch, E – total freshwater
habitat edge length. All metrics were calculated using FRAGSTATS and QGIS software and the resolution
of analysis and connectivity value were the same for all indicators.
Table 3: Landscape Metrics and their Mathematical Descriptions
Metric Formula Interpretation
Number of patches (NP) 𝑁𝑃 = 𝑁 Total identified habitat patches
Patch density (PD) 𝑃𝐷 = 𝑁𝐴 × 100 Patches per 100 ha
Mean patch area (AREAMN) AREAMN = 𝐴𝑇𝑁 Average habitat patch size
Edge density (ED) 𝐸𝐷 = 𝐸𝐴 Habitat boundary per unit landscape
Largest patch index (LPI) 𝐿𝑃𝐼 = 𝐴max
𝐴𝑇
× 100 Dominance of the largest patch
Landscape Resistance Surface
Resistance surfaces are used to show how hard it is to move through each land cover category.
Resistance values from 1 to 100 were assigned to five categories (see table 4). Freshwater and natural
vegetation had the lowest resistance while built -up areas had the highest resistance. These values are
modelling assumptions that have to be calibrated according to the target organism or process. Hydrological
channels and barriers need to be taken into account when dealing with aquatic animals (Johnson & Host,
2010). If there are multiple spatial variables involved, the resulting resistance value is calculated using
equation (3) below, where 𝑤𝑖 is the weight of variable 𝑖.
𝑅(𝑥) = ∑ 𝑤𝑖𝑅𝑖(𝑥)𝑛𝑖=1 (3)
Table 4: Landscape Resistance Values Assigned to Land-Cover Classes
Land-cover category Resistance value Resistance level
Freshwater bodies 1 Very low
Natural vegetation 5 Low
Agricultural land 30 Moderate
Barren/open land 50 High
Built-up area 100 Very high

Aquatic Ecosystems and Environmental Frontiers, Volume 4, Issue 4, December 2026 ISSN: 3049-186X
DOI: https://doi.org/10.70102/AEEF/V4I4/AF260431
www.aquaticfrontiers.com 266
III. Results and Analysis
Spatial Distribution of Freshwater Habitats
The spatial analysis resulted in the detection of 24 patches of freshwater habitat covering a total habitat
area of 3.84 km² (Table 5). It is seen from the mean patch area of 0.16 km² that the habitat is composed of
many small patches of habitat instead of a large continuous aquatic habitat. The biggest patch occupies an
area of 0.62 km² which constitutes 16.15% of the total habitat area, showing the lack of domination of any
water body within the habitat.
Table 5: Spatial Characteristics of Freshwater Habitat Patches
Parameter Result
Total study area 100.00 km²
Number of freshwater patches 24
Total freshwater habitat area 3.84 km²
Mean patch area 0.16 km²
Largest patch area 0.62 km²
Smallest patch area 0.04 km²
Freshwater habitat coverage 3.84%
Patch density 0.24 patches per 100 ha
Largest patch index 16.15%
Landscape Fragmentation Assessment
The landscape fragmentation was analyzed based on patch configuration and connectivity categories
(Table 6). Among the total of 24 patches, 14 (58.33%) were isolated, 7 (29.17%) were moderately
connected, and only 3 (12.50%) were highly connected. This ind icates that the freshwater landscape is
fragmented, and most of the habitat patches do not have any connection to their neighbors. Isolation of such
patches makes it essential to determine the intermediate habitat which can serve as connecting corridors,
especially when there is agricultural and developed land between such disconnected patches.
Table 6: Classification of Freshwater Habitat Connectivity
Connectivity category Number of patches Percentage
Isolated habitats 14 58.33
Moderately connected habitats 7 29.17
Highly connected habitats 3 12.50
Total 24 100.00
Landscape Connectivity Indicators
Table 7: Landscape Fragmentation and Connectivity Indicators
Indicator Baseline value Ecological interpretation
Number of habitat patches 24 Spatial distribution of habitats
Mean patch area 0.16 km² Relatively small habitat units
Largest patch index 16.15% Limited dominance of one patch
Patch density 0.24 per 100 ha Habitat patch frequency
Edge density 0.186 km/km² Habitat boundary characteristics
Habitat coverage 3.84% Freshwater proportion of landscape
Mapped corridor links 9 Modelled connections
Isolated habitat proportion 58.33% High spatial isolation

Aquatic Ecosystems and Environmental Frontiers, Volume 4, Issue 4, December 2026 ISSN: 3049-186X
DOI: https://doi.org/10.70102/AEEF/V4I4/AF260431
www.aquaticfrontiers.com 267
Table 7 presents a summary of the fragmentation and connectivity indices. There were 24 patches with
an average size of 0.16 km² and patch density of 0.24 per 100 hectares; the landscape is thus characterized
by numerous and relatively small habitat fragments. The edge density index was found to be 0.186 km per
km², while the area covered by water features was 3.84%. There were nine corridor connections between
patches, and 58.33% of the habitat fragments remained isolated from others.
Ecological Corridor Identification
The least-cost paths analysis on the resistance surface provided nine possible corridors (Figure. 2(a&b)).
Three were high-priority, four were moderate, and two were low -priority corridors. The shortest corridor
among these was C6 with a length of 0.80 km and cumulative resistance of 12.8, while the longest corridor
was C8 with length 2.75 km and cost 43.2, which implies that the landscape within this corridor is more
difficult for movement. The cumulative cost is not an actual distance.

Figure 2(a&b): Least-Cost Ecological Corridors Over the Landscape Resistance Surface (A) And
Corridor Cost and Length (B)
Habitat Connectivity
Table 8: Graph-Based Connectivity Characteristics of Selected Habitat Patches
Habitat patch Degree centrality Betweenness centrality Connectivity role
H01 3 0.42 Major connector
H03 4 0.68 Critical connector
H06 2 0.21 Intermediate habitat
H08 3 0.51 Major connector
H10 4 0.75 Critical connector
H14 2 0.18 Intermediate habitat
H16 3 0.39 Supporting habitat
H19 1 0.08 Peripheral habitat
H22 2 0.14 Supporting habitat
The contribution of individual habitats to the network was evaluated using graph analysis (Table 8).
H10 and H03 were the two habitats with highest betweenness centrality values, 0.75 and 0.68 respectively,
having four graph links in both cases, which were termed as critical connectors. H08 and H01 were
considered as major connectors while H19 was having the least value, 0.08. It means that the removal of
the critical connectors would lead to reduction in connectivity of the habitat network.

Aquatic Ecosystems and Environmental Frontiers, Volume 4, Issue 4, December 2026 ISSN: 3049-186X
DOI: https://doi.org/10.70102/AEEF/V4I4/AF260431
www.aquaticfrontiers.com 268
IV. Discussion
Fragmentation of the landscape can have an impact on the connection of freshwater ecosystems through
altering the extent and the spatial configuration of habitat areas. The identification of 14 separated patches
suggests high ecological isolation due to the expansion of agriculture, urban development and disruption of
drainages. The presence of nine ecological corridors shows that there are possible connections between the
patches using natural vegetation and riparian zones; but limitations such as seasonal water, artificial
obstacles and disconnection of the water flows could limit the movements of aquatic organisms. Hence,
field studies should be conducted to confirm the functionality of these connections.
In this research, the application of graph -based connectivity analysis revealed that H10 and H03 were
considered to be key habitat connectors as these habitat areas are maintaining ecological connections
between isolated freshwater ecosystems. In addition to graph centralities, the integration of habitat area,
landscape resistance, and connectivity measures increases the value of this approach for prioritizing
conservation efforts in a better way through identifying strategic habitats. The proposed framewor k
supports habitat protection, corridor restoration, barrier management, sustainable land -use planning, and
long-term ecological monitoring. Compared with individual assessment approaches, this integrated spatial
framework provides a comprehensive basis for freshwater conservation decisions.
V. Conclusion
This research paper proposes a landscape -based approach for assessing connectivity between isolated
freshwater ecosystems through remote sensing, GIS, landscape metrics, least -cost modeling and graph
analysis. Analysis of the 24 habitats resulted in the fr agmented network with 58.33% of isolated habitats,
as well as in mapping of the nine corridors that would connect them. Two habitats, H03 and H10, are crucial
for connection, while restoration of a hypothetical landscape increased the Probability of Connectivity from
18.60% to 31.40%. This approach provides evidence for the need for corridors’ restoration, habitat
protection and sustainable landscape management. There are several aspects that have to be considered.
Firstly, satellite image waterbody identification is sensitive to changes in water level and presence of clouds.
Secondly, resistance values will be species specific and require ecological validation. Thirdly, least -cost
paths show potential connections but not actual ones. Fourthly, hydrological barriers and seasonal
connectivity require further field research. Finally, results of connectivity analysis on graph will be
influenced by the chosen dispersal thresholds and other network assumptions and sensitivity analysis can
be conducted. Future rese arch could combine the use of machine learning based classification of water
bodies, species distribution modeling, hydrologic network modeling and circuit theory to determine
alternative routes for movement. Field survey-based observation of biodiversity, seasonal hydrological
connectivity, and specific dispersal information for species could make the estimations more robust. Future
long-term monitoring based on multi-seasonal satellite imagery and ecological movement could be used to
evaluate freshwater connectivity corridors.
References
[1] Erős, T., Olden, J. D., Schick, R. S., Schmera, D., & Fortin, M. -J. (2012). Characterizing
connectivity relationships in freshwaters using patch -based graphs. Landscape ecology , 27(2),
303-317. https://doi.org/10.1007/s10980-011-9659-2
[2] Luque, S., Saura, S., & Fortin, M. J. (2012). Landscape connectivity analysis for conservation:
insights from combining new methods with ecological and genetic data. Landscape ecology, 27(2),
153-157. https://doi.org/10.1007/s10980-011-9700-5
[3] Matos, C., Petrovan , S. O., Wheeler, P. M., & Ward, A. I. (2019). Landscape connectivity and
spatial prioritization in an urbanising world: A network analysis approach for a threatened
amphibian. Biological Conservation, 237, 238-247. https://doi.org/10.1016/j.biocon.2019.06.035

Aquatic Ecosystems and Environmental Frontiers, Volume 4, Issue 4, December 2026 ISSN: 3049-186X
DOI: https://doi.org/10.70102/AEEF/V4I4/AF260431
www.aquaticfrontiers.com 269
[4] Johnson, L. B., & Host, G. E. (2010). Recent developments in landscape approaches for the study
of aquatic ecosystems. Journal of the North American Benthological Society, 29(1), 41-66.
https://doi.org/10.1899/09-030.1
[5] Bishop-Taylor, R., Tulbure, M. G., & Broich, M. (2015). Surface water network structure,
landscape resistance to movement and flooding vital for maintaining ecological connectivity across
Australia’s largest river basin. Landscape Ecology, 30(10), 2045-2065.
https://doi.org/10.1007/s10980-015-0230-4
[6] McCullough, I. M., King, K. B., Stachelek, J., Diaz, J., Soranno, P. A., & Cheruvelil, K. S. (2019).
Applying the patch -matrix model to lakes: a connectivity -based conservation framework.
Landscape Ecology, 34(11), 2703-2718. https://doi.org/10.1007/s10980-019-00915-7
[7] Phillipsen, I. C., Kirk, E. H., Bogan, M. T., Mims, M. C., Olden, J. D., & Lytle, D. A. (2015).
Dispersal ability and habitat requirements determine landscape‐level genetic patterns in desert
aquatic insects. Molecular Ecology, 24(1), 54-69. https://doi.org/10.1111/mec.13003
[8] Ding, G., Guo, J., Yi, D., Ou, M., & Yang, G. (2025). Evaluating habitat isolation driven by future
urban growth: A landscape connectivity perspective. Environmental Impact Assessment Review ,
113, 107886. https://doi.org/10.1016/j.eiar.2025.107886
[9] Field, R. D., & Parrott, L. (2022). Mapping the functional connectivity of ecosystem services
supply across a regional landscape. eLife, 11, e69395. https://doi.org/10.7554/eLife.69395
[10] Miao, Z., Pan, L., Wang, Q., Chen, P., Yan, C., & Liu, L. (2019). Research on urban ecological
network under the threat of road networks—A case study of Wuhan. ISPRS International Journal
of Geo-Information, 8(8), 342. https://doi.org/10.3390/ijgi8080342

论文原文与出版信息

Aquatic Ecosystems and Environmental Frontiers
查看论文原文

继续探索