Linking landscapes: movement ecology and conservation priorities for the critically endangered lesser florican in India

1 Introduction

The Lesser florican (Sypheotides indicus) is critically endangered and the world’s smallest bustard species (BirdLife International, 2021). Endemic to the Indian subcontinent, it was historically widely distributed, with breeding areas in the Nashik, Ahilyanagar (old name is Ahmednagar), and Solapur districts of Maharashtra, eastern Haryana, and the Kathiawar Peninsula, and grassland patches of western Gujarat, including Kutch (Goriup and Karpowicz, 1985; Gadhvi and Shah, 2008; Bhardwaj et al., 2011). Due to habitat loss, this species now breeds in a few isolated grassland patches in Gujarat, Rajasthan, Maharashtra and Madhya Pradesh (Sankaran, 1991, 2000; Bhardwaj et al., 2011; Narwade et al., 2015). A recent survey, encompassing most of the known breeding areas in western Madhya Pradesh, southern Rajasthan, central and western Gujarat, and Maharashtra, documented an ~80% decline in the Lesser florican population over the last three to four generations (Dutta et al., 2018). The country wide survey estimated 340 territorial males (conservatively 240–264), corresponding to a total population of approximately 480 birds (assuming an equal sex ratio). This represents a drastic ~80% decline from an earlier estimate of ~3,530 birds in 1999 (Sankaran, 1999). Within this context, Blackbuck National Park (BNP) in Bhavnagar district of the Saurashtra region stands out as one of the key strongholds for the Lesser florican, supporting an estimated 95–115 males (Dutta et al., 2018). As a local migrant in India, the species’ movements are closely tied to seasonal rainfall patterns, and males exhibit moderate fidelity to their breeding sites (Ali and Ripley, 1969; Dharmakumarsinhji, 1950).

Grasslands that receive good rainfall, have moderately tall grass, and provide ample ground cover are the preferred breeding habitats of Lesser florican (Sankaran and Rahmani, 1986; Sankaran, 1999; Dutta et al., 2018). The use of croplands for breeding is also reported, though not common (Sankaran, 2000). In addition, low-grazing-intensity grasslands that are away from disturbance are additional predictors of their occurrence (Dutta and Jhala, 2014). In highly grazed grasslands, they tend to occur more frequently in croplands, although agricultural growth has a negative effect on the Lesser florican numbers during the breeding season (Sankaran, 1997). This preference for multi-use agro-grassland systems is believed to reflect the depletion and degradation of grasslands across the species’ range.

Grasslands are among India’s most threatened ecosystems, and less than 5% of its grasslands fall within protected areas (MadhuSudan and Vanak, 2023). These grasslands harbour several endemic and threatened species, including the great Indian bustard (Ardeotis nigriceps), the Lesser florican (Sypheotides indicus), and the blackbuck (Antilope cervicapra). At the same time, these ecosystems also support the livelihoods of several million nomadic pastoralists and their indigenous livestock (MadhuSudan and Vanak, 2023). The effective conservation of the Lesser florican thus requires an understanding of the species’ habitat requirements, movement patterns, socio-economic attributes and disturbance tolerance.

Western Gujarat, particularly Saurashtra and Kutch, part of the current study area, is an important conservation landscape due to the presence of a wide variety of threatened taxa and a long history of human-wildlife coexistence (Dutta and Jhala, 2014). Local agro-pastoral systems have historically been attuned to the region’s semi-arid and highly variable climate. Recent agricultural expansion has converted many of the region’s grasslands to cropland, intensifying grazing pressure on the remaining areas and likely having negative impacts on wildlife (Gadhavi et al., 2012; Dutta et al., 2013). Grassland-dependent species are thus facing the combined effects of natural habitat loss to agriculture and altered grazing practices on remnant grasslands (Dutta and Jhala, 2014; Gadhavi, 2020).

Historically, large tracts of semi-arid grasslands in India have been converted to croplands, resulting in substantial loss and fragmentation of suitable habitat (Tian et al., 2014; MadhuSudan and Vanak, 2023). Such land-use changes reduce the availability of structurally suitable grasslands required for breeding displays, nesting, roosting, and foraging, while also increasing disturbance from farming activities and associated infrastructure (Sankaran, 1997; Dutta and Jhala, 2014). In agro-grassland mosaics, floricans may still occur but typically at much lower densities, reflecting reduced habitat quality and increased fragmentation compared with contiguous grassland landscapes (Wolff et al., 2002; Dutta et al., 2018). Shifts in cropping patterns, intensification and mechanisation of agriculture, expansion of irrigation, and rapid shift to outsourced farming through contracts can rapidly transform these landscapes. This juxtaposition often leads to the loss or alteration of seasonal fallow lands and grass-dominated patches that are crucial for the Lesser floricans. As a result, even where agro-grassland mosaics currently provide some habitat value, they may represent transient or unreliable refugia rather than secure habitats in the long-term. Continued intensification of agriculture may not only further degrade remaining grassland habitats but also disrupt seasonal movement pathways, posing long-term challenges to the conservation of this grassland specialist.

In this study, we aimed to understand how Lesser floricans use space across seasons and behavioural contexts by combining resource selection analyses with movement metrics. Using resource selection functions (RSFs), we quantified habitat selection across seasons and between migratory strategies. We hypothesised that floricans would show strong positive selection for open grassland habitats during the breeding season, reflecting their well-known dependence on grasslands for display and nesting, and more flexible or opportunistic habitat use during the non-breeding season. The recent records from western Indian landscapes and telemetry observations indicate that individuals may occasionally traverse or utilise more heterogeneous habitats during local migration (Sivakumar et al., 2016; Ram et al., 2023). We therefore tested whether the tracked migrants and residents differ in their non-breeding habitat selection, predicting that tracked migrant individuals would exhibit broader or altered habitat associations owing to different space-use strategies and constraints imposed by unfamiliar landscapes.

In parallel, we also examined fine-scale movement patterns to assess how activity levels vary through time. We hypothesised that floricans would exhibit longer step lengths and higher movement speeds during the monsoon season than in other seasons, consistent with breeding-related activities such as reaching breeding grounds and finding suitable breeding areas or territories. Since the Lesser florican’s occurrence during the breeding season is tightly linked to local rainfall patterns (Sankaran, 1997; Dutta et al., 2018), which are unpredictable, we hypothesised that the larger step lengths and movement steps occur during the monsoon. We further tested whether movement rates differ between daytime and nighttime across seasons, and whether the strength of this day-night contrast changes seasonally. Specifically, we predicted that the difference would decrease during the post-monsoon period, potentially indicating a shift toward crepuscular activity following the breeding season.

2 Materials and methods

2.1 Study area

Twelve Lesser florican individuals were fitted with GPS-GSM transmitters at two primary sites in Gujarat, India, separated by approximately 450 km: (1) Blackbuck National Park (BNP) in Bhavnagar district of the Saurashtra region, and (2) Kutch Bustard Sanctuary (KBS) together with the adjacent Naliya grasslands in the Kutch district (Figure 1).

Blackbuck National Park (BNP) (21°56′ N, 72°10′ E) encompasses ~34.08 km² of semi-arid grassland and scrubland habitat, bordered by agricultural lands to the north and the Gulf of Cambay to the south. The site experiences winter temperatures of 1–38 °C, summer temperatures of 37–48 °C, and mean annual precipitation of ~468–518 mm, predominantly during the southwest monsoon (Gadhvi, 2003; Jethva and Jhala, 2004). Grasslands constitute ~60.9% of the area, dominated by Dicanthium annulatum and various Sporobolus spp., with associated shrub and tree cover including Acacia nilotica, Ziziphus, Salvadora, and Capparis (Figure 2). BNP is an Important Bird Area (IBA) falling under the semi-arid Gujarat-Rajwada biotic province of the semi-arid biogeographic zone (Rodgers et al., 2000). During the monsoon season, BNP has the highest concentration of Lesser florican individuals in India.

Kutch Bustard Sanctuary (KBS) (2.08 km²) and Naliya Grasslands (~15.93 km²) occur within a broader landscape of flat terrain characterised by Northern Tropical Thorn vegetation, including grasslands (Dicanthium, Aristida, Cymbopogon, Chrysopogon), interspersed with scrub (Acacia, Capparis, Ziziphus, Neltuma) and rain-fed croplands (groundnut, cotton, millet). The area experiences minimum temperatures of ~5 °C, summer maxima of 40–45 °C, and mean annual rainfall of ~384 mm. Naliya Grasslands, designated an Important Bird Area (Criteria A1), is the only site in India where sympatric populations of three bustard species occur: the critically endangered great Indian bustard, the vulnerable Asian houbara (Chlamydotis macqueenii), and the Lesser florican. These two sites are an arid mosaic of grasslands and scrub, modified by rain-fed croplands and pastoralism, and represent one of the crucial habitats for the conservation of this critically endangered species in Gujarat (Figure 3). The dominance of Dicanthium, Cymbopogon, Sporobolus, Sehima and Chrysopogon is especially significant because these species create the ideal structural conditions for the Lesser florican, offering suitable height for nesting concealment, open display areas for males, and support for a rich invertebrate food base. Their vigorous growth during good rainfall years provides the tall grassland structure (>1.5 m) the species prefers, while still remaining usable under moderate grazing pressure. This makes them critical for breeding habitat quality and overall conservation of this critically endangered species.

2.2 Bird trapping and transmitter deployment

Lesser floricans were captured and tagged for telemetry in two phases. In 2020, one male and one female were tagged at BNP. During the second phase (2021–2022), 10 additional males were tagged, bringing the total sample to 12 individuals (Supplementary Table 1). Breeding males were easy to detect and capture because they perform conspicuous courtship displays from fixed display sites within a larger lek in their black breeding plumage. In contrast, females remain camouflaged and secretive, which greatly reduces their detectability. To avoid compromising breeding success in a critically endangered species, we did not attempt to trap nesting females. Eight birds tagged in BNP were deployed with solar-powered GPS-GSM transmitters mounted as backpacks, and among the four birds tagged in and around KBS, three were deployed with PTT transmitters and one with a GPS-GSM transmitter (Supplementary Table 1). Tags recovered from dead birds (Ram et al., 2022) were reused during the second phase. Three transmitters (12 g; Microwave Telemetry, Columbia, Maryland, USA) transmitted data via the Argos satellite system, whereas nine transmitters (10 g; Ornitela, UAB, Vilnius, Lithuania) used the GSM cellular network. Birds were captured by experienced trappers following protocols developed in earlier works (Ram et al., 2022, 2023). In all cases, tag mass was less than 3% of the body mass of the tagged birds (Supplementary Table 1).

2.3 Spatial extent of space use and availability for habitat selection

A total of 160,166 locations were recorded for all Lesser floricans. For each individual, locations were recorded as latitude–longitude coordinates with associated timestamps. Timestamps were standardised to account for variation in date-time formats among individuals, and records with missing coordinates or unparseable timestamps were removed. This ensured that only fixes with reliable spatial and temporal information were retained for analysis. To place each location in its geographic context, all points were initially handled in a geographic coordinate system (WGS84, EPSG:4326) and spatially joined to a polygon layer representing the boundaries of the Indian Republic. This step was used to verify spatial consistency and to support exploratory summaries of movements across states. For all subsequent spatial analyses, locations were projected to UTM Zone 42N (EPSG:32642), which is appropriate for the study region and allows distance and area-based calculations in metres.

2.4 Seasonal classification and habitat variables

Each GPS location was assigned to one of two biologically meaningful seasons based on the month of the year. We defined the breeding season as the monsoon period (June–September), which corresponds to the known breeding period of Lesser floricans in Gujarat. All remaining months (October–May) were classified as the non-breeding season. This seasonal grouping allowed us to quantify and compare habitat selection between breeding and non-breeding periods.

Habitat information was derived from the ESA World Cover 2020 land-use/land-cover (LULC) product, which provides a global land-cover map at a 10 m spatial resolution based on Sentinel-1 and Sentinel-2 imagery and is aligned with the UN-FAO Land Cover Classification System (Zanaga et al., 2021). Within our study area, nine LULC classes were present: built-up, cropland, grassland, herbaceous wetlands, mangroves, shrubland, sparse vegetation, tree cover, and water bodies. For both used and available locations, the corresponding LULC class was extracted and used as a categorical habitat predictor in subsequent analyses. Grassland was chosen as the reference habitat class because it represents the species’ core and most characteristic habitat. Although ESA WorldCover 2020 provides a consistent land-cover product across the study region, some local land-use changes may have occurred during 2021–2025. However, the broad habitat categories used in this study (grassland, cropland, shrubland, tree cover, wetlands, built-up areas, and water) remained largely stable across the study landscapes, and field observations indicated no major landscape transformations likely to alter the principal ecological interpretations. In addition, to evaluate the influence of artificial night-time light on habitat use, we extended the resource selection function (RSF) framework to include night-time light intensity as a continuous environmental covariate. Artificial night-time light was quantified using the global annual simulated VIIRS night-time light dataset (1992–2023; Chen et al., 2024). The annual simulated VIIRS night-time light dataset is available through 2023. Accordingly, telemetry locations recorded during 2024–2025 were assigned values from the 2023 layer, which represented the most recent available estimate of anthropogenic night-time illumination. Only two transmitters remained active beyond 2023, with one ceasing transmission on 26 November 2024 and the other on 3 January 2025.GPS timestamps were converted from Coordinated Universal Time (UTC) to local time (India Standard Time, UTC + 5:30), and fixes recorded between 18:00 and 06:00 were classified as nocturnal. Only nocturnal locations were retained for these analyses, as artificial lighting is ecologically relevant primarily at night.

2.5 Defining seasonal availability

To define habitat availability in a biologically realistic way, we used individual and season-specific availability domains. For each individual in each season (breeding and non-breeding), we constructed a 100% Minimum Convex Polygon (MCP) using all locations available for that individual–season combination. MCPs were generated only when at least 30 fixes were available to ensure a minimum level of spatial representation. The 100% MCP approach was chosen for several reasons. First, it provides a simple, transparent estimate of the area potentially accessible to each individual in each season. Second, because the primary aim of this study is conservation-oriented: namely, to identify areas potentially used by Lesser floricans across breeding and non-breeding ranges, excluding locations using movement-based or behavioural filters could have artificially truncated the species’ spatial extent and led to an underestimation of areas requiring protection. Retaining all locations, therefore, offers a more precautionary and inclusive representation of space use. Third, the 100% MCP approach avoids the convergence and stability issues encountered with kernel-based home-range estimators in this dataset. Finally, defining availability at the individual × season level reduces bias that would arise from pooling individuals with very different movement strategies. The area of each MCP was calculated and summarised to provide descriptive statistics on seasonal space use.

2.6 Resource selection function framework

We used a Resource Selection Function (RSF) framework to relate Lesser florican space use to habitat and human influence in a manner directly relevant to conservation and management. All observed GPS locations were treated as “used” points. For each individual × season MCP, we generated random “available” points by uniformly sampling within the MCP, using a 5:1 ratio of available to used points to obtain stable estimates without inflating the sample size. Used and available points were combined into a single dataset with a binary response (1 = used, 0 = available), and each point retained information on individual identity, season, and habitat class.

Habitat selection was modelled using logistic regression RSFs, with LULC included as a categorical predictor and grassland as the reference category, reflecting its role as the species’ core habitat. Individual identity was included as a random intercept in mixed-effects models to account for repeated measures and differences in overall space-use intensity among birds. Model coefficients were interpreted as log-odds of use relative to availability and were exponentiated to odds ratios for clearer ecological interpretation.

We fitted three sets of models to address conservation-relevant questions about when and where Lesser floricans use different habitats. First, to test whether habitat selection changes between seasons, we fitted a mixed-effects model with a habitat × season interaction [used ~ LULC × season + (1 | ID)]. This model tests the hypothesis that selection strength and direction differ between the breeding and non-breeding periods, a finding that is critical for identifying season-specific conservation priorities. For interpretation and visualisation, we also fitted season-specific models [used ~ LULC + (1 | ID)] for each season.

Second, based on movement behaviour, individuals were classified a priori as migrants, Lesser florican male-9 (LFM9), Lesser florican male-10 (LFM10), and Lesser florican female (LFF) or residents (all others). Because migratory behaviour is expressed outside the breeding season, all migrant–resident comparisons were restricted to the non-breeding period.

Third, to test the hypothesis that tracked migrants and residents differ in habitat selection during the non-breeding season, we fitted an interaction model [used ~ LULC × migrant + (1 | ID)]. This model formally evaluates whether the two groups differ in their use of each habitat class relative to grassland. For clearer presentation, we also fitted separate models for residents (mixed-effects) and three tracked migrant individuals (simple logistic regression, due to small sample size), which were used to derive group-specific effect sizes and odds ratios, while the interaction model provided the basis for inference. Owing to the small number of tracked migratory individuals, the migrant analysis was treated as exploratory and fitted using a fixed-effects logistic regression.

Because exposure to human-modified landscapes is likely to be greatest during the non-breeding season, we also examined the influence of light pollution as a proxy for human pressure. In this study, we treated night-time light as a broader gradient of anthropogenic landscape modification. We fitted a mixed-effects RSF including habitat, standardised night-time light intensity (NTL), migratory strategy, and their interactions [used ~ LULC × migrant + NTL × migrant + (1 | ID)]. Here, the main effect of NTL represents residents’ response, while the interaction term quantifies how tracked migrants differ. Given the small number of tracked migrants and sparse use of some habitats, inference focused primarily on the effects of NTL and its interaction with migratory strategy. All models were fitted using maximum likelihood in the lme4 package (Bates et al., 2015), and effect sizes are reported with 95% confidence intervals. Across all models, availability was defined within individual seasonal home ranges (100% MCPs), and the resulting used–available models were interpreted as estimates of relative habitat selection within the accessible landscape. To evaluate whether habitat-selection results were sensitive to the definition of habitat availability, we repeated the global RSF using three used-to-available sampling ratios (1:2, 1:5, and 1:10) while retaining the same individual × season 100% minimum convex polygons (MCPs) to define available habitat. Parameter estimates, confidence intervals, and the direction and relative magnitude of habitat-selection coefficients were compared across models to assess the robustness of habitat-selection inference.

2.7 Landscape context

We initially explored landscape configuration using moving-window metrics at multiple spatial scales, but these measures showed little variation and frequent missing values due to the extreme fragmentation and small size of grassland patches in the study area. As a result, patch- and class-level configuration metrics were not informative for point-level selection analyses and were excluded from the final RSFs. Instead, we used the Euclidean distance to the nearest grassland cell as a simpler, more robust measure of landscape context with direct conservation relevance. This variable was calculated on a binary grassland raster, extracted at each point location, and scaled prior to modelling. Distance to grassland provides an interpretable gradient of proximity to core habitat and allows for assessing how space use changes as birds move farther from remaining grassland patches, information directly useful for guiding habitat protection and restoration efforts.

2.8 Seasonal movement patterns

To describe how Lesser floricans move across the landscape, we quantified step lengths (great-circle distances between consecutive locations) and average speeds (step lengths divided by the time intervals between fixes) using the Haversine formula. To avoid spurious values, we retained only steps with positive distances and speeds < 70 km h⁻¹, resulting in a final dataset of 157,300 steps from 12 individuals (11 males and 1 female).

We then tested whether movement patterns varied across seasons (summer: March–May; monsoon: June–September; post-monsoon: October–November; winter: December–February) and between day (06:00–17:59 IST) and night (18:00–05:59 IST), as these differences are directly relevant for understanding seasonal space-use strategies and identifying periods of heightened activity or risk. We fitted separate linear mixed-effects models for step length and speed, with season, day–night period, and their interaction as fixed effects, and individual identity as a random intercept to account for repeated measurements [log(response) ~ season × day_night + (1 | ID)]. Both response variables were log-transformed to meet the assumptions of normality.

Overall effects were assessed using Type III Wald χ² tests, and Tukey-adjusted pairwise comparisons of estimated marginal means were used to explore differences among seasons and between day and night. All analyses were conducted in R (R Core Team, 2025) using the lme4 (Bates et al., 2015), lmerTest (Kuznetsova et al., 2017), and emmeans (Lenth and Piaskowski, 2025) packages. This approach allowed us to quantify when Lesser floricans move most and how movement intensity varies across the annual cycle, providing insights directly relevant to the timing of conservation actions and the mitigation of seasonal threats.

2.9 Recurrent migration corridors

Telemetry derived migration corridors were reconstructed using a smoothed line-density approach (Sahlean et al., 2020; Tian et al., 2022) to quantify spatial overlap among migration trajectories. We converted the GPS locations into movement trajectories for each individual and migration cycle, then rasterised them onto a 5 km grid spanning all tracks. Each cell recorded the number of intersecting trajectories, thereby mapping the pathway-use intensity. To handle positional uncertainty and natural route variation, we applied spatial smoothing using a 5×5-cell moving-window kernel (~25 km radius). The resulting density surface was normalised to 0–1, indicating relative migration probability across the landscape.

The high-density migration corridor was defined as areas where the normalised flyway probability exceeded 0.25. This threshold identified regions of consistently repeated use across individuals and migration cycles while retaining biologically meaningful spatial continuity. The sensitivity of corridor delineation to the smoothing scale was evaluated using kernel sizes of 5, 7, and 9 cells. Core corridor area estimates were compared across kernels to confirm the robustness of spatial inference. To quantify the spatial structure of migration, the high-density corridor raster was converted to a polygon representation. Regular transects spaced at 20-km intervals were generated perpendicular to the longitudinal extent of the corridor. Each transect was intersected with the corridor polygon, and the corridor width was calculated as the total length of the intersection segments. This produced a continuous width profile describing variation in corridor breadth along the migration axis. Migration flexibility was defined as the coefficient of variation of corridor width and quantified as the Migration flexibility index (MFI):

MFI=SD(width)/Mean(width)

Higher values indicate diffuse migration pathways with greater spatial freedom, whereas low values indicate constrained funnel-like migration.

Habitat composition data (LULC) were used to evaluate landscape influence on corridor structure. Each transect was buffered by 10 km to represent the ecological landscape context influencing movement decisions. Habitat composition within buffers was extracted and converted into proportional coverage of each habitat class. Grassland habitat was treated as the ecological reference category, and the relationship between corridor width and habitat composition was evaluated using multiple linear regression.

2.10 Above-ground level flight

We estimated flight height above ground level (AGL) from 9 Lesser florican individuals equipped with Ornitela GSM/GPS tags. Argos Doppler PTT data for three individuals (LFF, LFM1, and LFM5) were excluded from quantitative AGL calculations due to unreliable altitude measurements (Poessel et al., 2018). Transmitter altitude (m) was subtracted from ground elevation extracted from a digital elevation model (DEM) at each location to calculate raw AGL (AGL_raw = transmitter altitude – DEM elevation). For GSM/GPS tags, which provide direct ellipsoidal altitudes with reasonable vertical accuracy, we applied an individual-level bias correction by subtracting the 5th percentile of raw AGL values as an estimate of systematic offset due to GPS vertical error and minor DEM inaccuracies, as proposed by Poessel et al. (2018). Extreme outliers were removed (AGL_offset < –50 m or > 500 m), and the final AGL was floored at 0 m to reflect biologically realistic heights for a ground-foraging Lesser florican. Negative raw AGL values, common in low-relief environments due to measurement error, were retained prior to correction rather than discarded to avoid biasing distributions toward higher flights (Poessel et al., 2018).

3 Results

One transmitter remains active and is still transmitting data at the time of writing this manuscript. All other transmitters have stopped functioning (Supplementary Table 2). Overall, transmitters remained active for 4466 days, with the highest number of days recorded for LFM10 at 1093 and the lowest for LFM7 at 27 (Supplementary Table 2). Of 160,166 recorded locations, 150,147 fixes were retained for analysis after duplicates were removed. Across the study period (5.5 years), Lesser floricans spent the most time in croplands (3.19 years), followed by grasslands (1.56 years), shrublands (0.39 years), and sparse vegetation habitats (0.32 years). In contrast, croplands accounted for approximately 56% of the available habitat in the study area, whereas grasslands accounted for only about 5%. Shrublands and sparse vegetation made up roughly 10% and 8% of the available habitat, respectively. The area of seasonal MCPs varied widely among individuals, reflecting substantial heterogeneity in space use. The average MCPs, for all individuals, were larger during the non-breeding season (80,542.37 ± 42,410.44 km²; mean ± SE) than during the breeding season (58,491.40 ± 29,653 km²; mean ± SE), consistent with broader movements outside the core breeding grounds. Lesser floricans ranged across seven districts in Gujarat (Figure 4), while the migrant Lesser floricans ranged across six states in India (Figure 5). During the study period (2020, 2021, and 2022), the Saurashtra and Kutch sub-divisions received an average rainfall of 1146, 627, and 534 mm, respectively (Mohanty and Lal, 2021; Srivastava et al., 2022).

3.1 Above-ground level flight patterns

After bias correction, outlier removal, and flooring at 0 m, flight height above ground level (AGL) was reliably estimated only for the 9 individuals equipped with Ornitela GSM/GPS tags. The weighted mean flight height across GSM-tagged individuals was 24.4 ± 17.9 m (SD; n = 142,389 fixes) (Figure 6). Density distributions showed a primary mode around 20 m and a secondary peak near 0 m, with a right-skewed tail indicating occasional flights up to 50–150 m in some individuals (LFM8, LFM9, and LFM10 exhibited the widest ranges and the highest outliers). Overall, corrected AGL values confirmed that Lesser floricans in this study rarely sustained flights above 50 m, except for brief bursts during migration.

3.2 Patterns of habitat selection

Habitat selection by Lesser floricans showed a clear and consistent dependence on both habitat type and season, with additional divergence between resident and migratory individuals during the non-breeding period. The global mixed-effects resource selection function (RSF), which included an interaction between habitat (LULC) and season, revealed strong habitat and season-specific differences in habitat selection (Supplementary Table 3). Across seasons, grassland (the reference habitat) was consistently the most strongly selected habitat. Most other habitat classes had negative coefficients relative to grassland, indicating lower use than expected based on availability. However, several significant habitat × season interaction terms showed that the strength of avoidance or selection varied between the breeding and non-breeding seasons, suggesting some seasonal flexibility around a fundamentally grassland-centred strategy.

3.3 Breeding and non-breeding season RSFs

During the breeding season, this pattern was especially pronounced. Residents showed a strong and consistent preference for grassland over all other habitat classes (Table 1). Every non-grassland habitat had a strongly negative coefficient relative to grassland, indicating avoidance. The strongest avoidance was observed for mangroves, water bodies, herbaceous wetlands, built-up areas, and tree cover, while sparse vegetation, shrubland, and cropland showed the weakest avoidance (though still negative) effect among non-grassland habitats (Figure 7). Together, these results indicate that during the breeding season, Lesser floricans behave as strict grassland specialists, making very limited use of alternative habitat types.

Table 1. Fixed-effect coefficients (log-odds) and 95% confidence intervals from the breeding-season RSF, with grassland as the reference habitat and individual included as a random intercept.

In the non-breeding season, the overall pattern remained qualitatively similar: grassland continued to be the most strongly selected habitat, and all other habitat classes still showed negative coefficients relative to grassland (Table 2). However, the magnitude of avoidance varied among habitat types and was generally less extreme in cropland, shrubland, and sparse vegetation than in herbaceous wetlands, tree cover, mangroves, built-up areas, and water bodies (Figure 8). This suggests that although grassland remains the core habitat outside the breeding season, there is some population-level expansion into alternative open habitats, particularly cropland and shrubland.

Table 2. Fixed-effect coefficients (log-odds) and 95% confidence intervals from the non-breeding-season RSF, with grassland as the reference habitat and individual included as a random intercept.

3.4 Migrant versus resident habitat selection in the non-breeding season

Based on movement behaviour, individuals were classified as residents only for the period during which transmitters were active, and no long-distance seasonal movements were observed; movements after transmitter failure could not be assessed. As post-transmitter movements could not be assessed, these classifications represent observed movement behaviour during the monitoring period and should not be interpreted as definitive lifetime movement strategies. The three individuals (LFM9, LFM10, and LFF) were classified as migrants, while the remaining individuals were classified as residents. All comparisons between these groups were restricted to the non-breeding season.

The mixed-effects RSF, including an interaction between habitat and migratory strategy, revealed strong and significant differences in habitat selection between the tracked migrants and residents (Supplementary Table 4). The main effect of the migratory strategy indicated that tracked migrants used the grassland substantially less than residents. Several habitat × migrant interaction terms were large and positive, particularly for cropland, shrubland, tree cover, and water bodies, indicating that tracked migrant individuals used these habitats much more frequently than residents relative to grassland. For some habitat classes (e.g., built-up areas, herbaceous wetlands, sparse vegetation), estimates were unstable and associated with very large standard errors, reflecting near-zero use of these habitats by migrants. Overall, these results indicate a marked, habitat-specific divergence in space use between migratory and resident individuals during the non-breeding season, with migrant individuals exhibiting a broader, more flexible habitat-use pattern. Habitat-selection coefficients were highly consistent across the three availability sampling ratios (1:2, 1:5, and 1:10; Supplementary Table 5). The direction, relative magnitude, and statistical significance of habitat-selection coefficients remained largely unchanged, with grassland consistently identified as the most strongly selected habitat. These results indicate that the principal ecological inferences were robust to alternative definitions of habitat availability.

3.5 Selection relative to distance from the grassland

In the non-breeding season, the RSF showed a highly significant negative effect of distance on habitat use (β = −6.47 ± 0.14 SE, z = −46.57, p < 2 × 10⁻¹⁶). This indicates that Lesser floricans strongly selected locations on or immediately adjacent to grassland cells, with the probability of use dropping sharply with increasing distance. The random intercept for individual identity had low variance (0.071, SD = 0.267), suggesting relatively consistent selection across individuals. The model exhibited a singular fit, which is expected when selection is extremely strong and uniform.

A similar pattern was observed during the breeding season. Distance to the nearest grassland again had a highly significant negative effect (β = −5.66 ± 0.15 SE, z = −38.24, p < 2 × 10⁻¹⁶), with even lower random-effect variance (0.018, SD = 0.136), indicating highly consistent selection across individuals. The large negative coefficients in both seasons correspond to extremely small odds ratios (0.0034 in the breeding season), meaning that even small increases in distance from grassland dramatically reduced the probability of habitat use. The singular fits in both models further indicate that this pattern was nearly uniform across all individuals. Together, these results are consistent with the ecology of Lesser floricans as grassland specialists in a highly fragmented landscape, where individuals concentrate their space use on small, remnant grassland patches.

3.6 Effects of night-time light pollution

An increase in nighttime light significantly reduced the odds of use among residents (β = -0.87, OR = 0.42, 95% CI = 0.40–0.44). Because nighttime light is closely associated with settlements, transportation networks, and other human infrastructure, it is interpreted here as an index of anthropogenic landscape modification rather than the independent effect of artificial light. The significant migrant × night-light interaction (β = 1.42, OR = 4.15, 95% CI = 2.63–6.56) indicates that tracked migrants responded differently to increasing light intensity, showing a substantially weaker positive response than residents. Combining the main and interaction effects yielded a weakly positive net association for tracked migrants (combined OR ≈ 1.74). Expressed as odds ratios, a one-standard-deviation increase in night-time light reduced the odds of use by approximately 58% for residents but increased the odds of use by approximately 73% for migrants. This highlights a pronounced divergence in nocturnal space use between resident and migratory individuals in relation to anthropogenic light pollution across its range, except for a few protected grassland patches. The migrant response to night-time light may partly reflect the scarcity of dark refugia along migration pathways, many of which are increasingly illuminated by anthropogenic activities. Thus, the observed pattern is unlikely to reflect an active preference for illuminated areas; rather, it may indicate a constraint imposed by widespread light pollution during migration, compared to non-migratory periods.

3.7 Seasonal movement patterns

Step lengths and movement speeds differed significantly among seasons, between day and night, and showed a significant season × day_night interaction (Table 3). Step lengths and speeds were highest during the monsoon and lowest during the post-monsoon and winter seasons (Table 4). Across all seasons, daytime movements were substantially longer and faster than nighttime movements (overall ~70–72% reduction at night). The day–night contrast was strongest in summer and winter and weakest in the post-monsoon season.

Table 3. Type-III analysis of deviance for linear mixed-effects models of log-transformed step length and speed.

Table 4. Estimated marginal mean step length and speed (± approximate 95% confidence intervals) by season and diel period.

Values are back-transformed from log-scale models.Bold values indicate that monsoon step lengths and speeds were significantly higher than all other seasons (P < 0.0001).

Pairwise seasonal contrasts (Tukey-adjusted) showed that monsoon step lengths and speeds were significantly higher than all other seasons (all P < 0.0001), while post-monsoon and winter did not differ from each other (P > 0.35). Day–night contrasts within seasons (all P < 0.0001) confirmed a strong reduction in movement at night in all seasons, with the smallest contrast in the post-monsoon (61–65% reduction) compared with summer and winter (71–74% reduction).

3.8 Individual migration trajectories and flyway structure

Three individuals exhibited local migratory movements during the study: a female (LFF) tagged in 2020 and two males (LFM9 and LFM10) tagged during the second phase of the study. The female dispersed from Gujarat to Telangana in November 2020 and remained there for approximately six months before her mortality during her return journey after colliding with a high-tension powerline in Maharashtra (Ram et al., 2022). Following earlier movements documented in Ram et al. (2023), the present study focuses on migration after November 2022 for the two remaining migratory males. These individuals displayed contrasting strategies: LFM10 followed a relatively structured seasonal circuit between western and southern India, repeatedly moving among Maharashtra, Karnataka, Gujarat, and Rajasthan, whereas LFM9 exhibited a more flexible trajectory, including a prolonged nomadic phase in Madhya Pradesh and a full monsoon stay in Gujarat (Supplementary Table 6). Over the four-year tracking period, LFM9 travelled a total of 3,036 km during migration, spending 196 days in transit and averaging 15.49 km per day (≈0.65 km per hour), while LFM10 travelled 4,873 km over 134 migratory days, averaging 36.36 km per day (≈1.52 km per hour).

Despite these individual differences in movement behaviour, migration trajectories from all tracked individuals formed a coherent north–south flyway extending across western and central India. Spatial smoothing of trajectory density revealed a broad migration pathway characterised by overlapping but partially parallel routes rather than a single concentrated track. The resulting flyway surface indicated repeated use of multiple spatially adjacent pathways across migration cycles and among individuals (Figure 9). A total of 14 migration cycles were analysed to construct the flyway surface: 8 from LFM10, 5 from LFM9, and 1 from LFF. These trajectories contributed variably to the density surface, with LFM10 and LFM9 being the primary contributors due to their multi-year tracking data. Sensitivity analyses with alternative smoothing kernels yielded comparable corridor extents, indicating that corridor delineation was robust to analytical parameter choices. The broader migration flyway (probability ≥ 0.05) covered approximately 206,322 km², while the high-density migration corridor (probability ≥ 0.25) occupied a substantially smaller area of 59,826.59 km², representing areas of repeated use across individuals. The mean corridor width was 70.8 ± 56.9 km. Corridor width ranged from relatively narrow segments to broad sections where migration routes diverged into multiple pathways. The Migration Flexibility Index, calculated as the coefficient of variation of corridor width, was 0.80, indicating exceptionally high variability relative to the mean corridor width. This high variability reflects both the spatial freedom within uniform landscapes and opportunistic searches for suitable habitat patches and foraging areas during migration. Among habitat variables, tree cover showed a significant negative relationship with corridor width (β = −111 km, p = 0.006); croplands exhibited a marginal negative effect on corridor width (β = −32 km, p = 0.064), suggesting moderate landscape resistance but not strong restriction of movement.

4 Discussion

Our study provides one of the most detailed movement-based assessments of space use, habitat selection, and seasonal behaviour in the Lesser florican within the context of rapidly transforming semi-arid agro-grassland landscapes. Of all the transmitters, one is still working, and the accumulated dataset (4,466 transmitter-days and 150,147 filtered GPS fixes) reveals consistent, biologically coherent patterns in habitat dependence, movement strategies, and responses to anthropogenic pressures. Together, these results reinforce the Lesser florican’s status as a grassland specialist while highlighting how agricultural expansion, habitat loss, fragmentation, and night lights are reshaping its realised niche. The current study area supports one of the largest remaining breeding populations (95–115 males; Dutta et al., 2018), within a broader, fragmented landscape where the species has declined by ~80% nationally (Dutta et al., 2018). These findings highlight the study area’s disproportionate importance for a critically endangered species now estimated at only ~480 individuals nationwide.

4.1 Grassland dependence and habitat selection in a fragmented landscape

Across seasons, grassland emerged as the most selected habitat. All other habitat classes were used less than expected relative to grassland, and this contrast was especially pronounced during the breeding season, when Lesser floricans behaved as near-obligate grassland specialists. Outside the breeding period, avoidance of cropland and shrubland was weaker, indicating some seasonal flexibility, but grassland nevertheless remained the core habitat. This pattern is consistent with previous ecological understanding of the species, which emphasises the importance of open grasslands with suitable vegetation structure for foraging, display, and nesting (Sankaran, 1997; Dutta and Jhala, 2014; Dutta et al., 2018). The strength of this dependence is most clearly illustrated by our distance-to-grassland models. In both breeding and non-breeding seasons, the probability of habitat use declined rapidly with increasing distance from the grassland, with even small increases in distance causing a dramatic drop in use. Such a sharp distance decay implies that Lesser floricans are effectively constrained to the immediate vicinity of grassland patches. From a landscape perspective, this means that not only the total area of grassland but also its spatial configuration, isolation, fragmentation, and connectivity are likely to be crucial determinants of space use and persistence (Wolff et al., 2002; Bretagnolle et al., 2024; Canonne et al., 2024).

Such a strong spatial association with grassland patches may also reflect a degree of site fidelity to historically suitable breeding habitats. Earlier behavioural observations and ringing studies of Lesser floricans indicated that individuals repeatedly returned to the same display areas across seasons (Dharmakumarsinhji, 1957), suggesting fidelity to traditional grassland sites. However, many of these landscapes have undergone extensive land-use change over the past century. Large areas of semi-arid grasslands in India have been converted to agriculture, with estimates suggesting a loss of approximately 20 million hectares of grasslands alongside major cropland expansion (Tian et al., 2014), while Open Natural Ecosystems (ONEs) continue to be among the most easily converted land-cover types in the country (MadhuSudan and Vanak, 2023). This conversion may not be conducive to Lesser floricans for performing important breeding behaviours, such as display, as observed in related bustard species. For instance, the great bustard (Otis tarda) display sites often occur in flat, open areas that maximise visibility and reduce predation risk, and these sites tend to remain consistent across years (Alonso et al., 2012). Similar behavioural tendencies in Lesser floricans may further reinforce their dependence on specific grassland patches, making them particularly vulnerable to the progressive conversion and fragmentation of traditional display landscapes.

4.2 Linking movement to land-use change

The movement-based results provide an important behavioural complement to landscape-scale assessments of the Lesser florican conservation status. The recent nationwide status assessment of the Lesser florican documents a severe population decline and a strong contraction of the breeding range, with birds now restricted to a small fraction of historically occupied sites (Dutta et al., 2018). Importantly, it also shows that floricans reach their highest densities in relatively small but contiguous grassland blocks, whereas in extensive agro-grassland mosaics they occur at much lower densities, spread thinly across large areas. This contrast implies that large, intact and well-managed grasslands are far more efficient for sustaining populations than vast but fragmented agricultural landscapes, as also observed in the little bustard (Tetrax tetrax) and other bustard species inhabiting open ecosystems (Wolff et al., 2002; Alonso et al., 2012; Faria and Morales, 2018).

Although Lesser floricans spent substantial time in croplands at the landscape scale, they did not treat croplands as a functional substitute for grassland. Instead, they remained spatially linked to grassland remnants and used the surrounding agricultural matrix only secondarily, particularly outside the breeding season (Sankaran, 1997; Mohan et al., 2015; Sivakumar et al., 2016; Collar, 2021; Ram et al., 2023). The extremely strong effect of distance from grassland, together with the large seasonal MCPs observed during the non-breeding period, points to a landscape in which suitable habitat is fragmented and embedded within a dominant agricultural matrix. In such settings, individuals are forced to range more widely to meet their needs, increasing movement costs and potentially exposure to threats, and leading to isolation among sub-populations (Wolff et al., 2001; Gray et al., 2007; Uddin et al., 2021).

This interpretation is fully consistent with landscape-scale analyses showing that continued agricultural expansion, intensification and mechanisation are likely to drive further declines in Lesser florican populations, primarily through the conversion and degradation of grasslands and the loss of structurally suitable habitat (Dutta et al., 2018). Our results suggest that even where low-intensity or traditional agriculture may offer some seasonal or opportunistic foraging opportunities, it cannot compensate for the loss of grassland itself (Kher and Dutta, 2021). Moreover, in the absence of suitable grassland or when it is scarce, they use such suboptimal habitats to meet their life requirements. Therefore, Lesser floricans do not simply shift into croplands as grasslands disappear; rather, they remain tightly bound to grassland remnants, making them acutely sensitive to both habitat loss and fragmentation.

Previous work has emphasised that the occurrence of Lesser florican is driven not only by land-cover identity (grassland versus cropland) but also by vegetation structure, particularly herbaceous biomass (50-100%) and ground-vegetation height (35–65 cm), which influence foraging efficiency, concealment, and visibility for display (Dutta and Jhala, 2014). This helps interpret the seasonal relaxation in avoidance of cropland and shrubland observed in our study. These habitats are likely used when their structure temporarily approximates suitable open conditions, rather than because they provide ecological functions equivalent to those of grasslands. During the breeding season, when display and nesting impose stricter structural requirements, Lesser floricans in our study showed the strongest and most exclusive selection for grassland, reinforcing the idea that true breeding habitat is both structurally and spatially constrained (Sankaran, 1999, 2000; Dutta and Jhala, 2012; Dutta et al., 2018; Ram et al., 2023). The unimodal relationship between herbaceous biomass and Lesser florican occurrence suggests a preference for intermediate conditions, where an optimal balance between grassland extent and vegetation height is maintained, rather than areas dominated by intensive agriculture or excessively tall grasslands (Dutta and Jhala, 2014).

Recent work on the Bengal florican (Houbaropsis bengalensis) demonstrates that community-managed grasslands can support viable populations of threatened bustards, even in landscapes subject to regular human use (Rahmani et al., 2017). Gray et al. (2007) showed that Bengal florican’s occurrence in the Tonlé Sap floodplain was more strongly related to vegetation structure maintained by traditional practices such as burning and grazing than to direct measures of human disturbance, providing empirical support for community-based conservation of grasslands. In the community-based conservation context, it is highly relevant in the Indian scenario (especially in areas with florican presence), where many highly degraded grassland areas remain as gauchar (community-owned grazing lands) or so-called wasteland in villages. Through grassland restoration efforts, these areas can serve as refuge patches for Lesser floricans and support local villages by increasing fodder availability and livestock security. This is highly relevant to the Lesser florican, which in our study showed strong and consistent selection for grassland and extreme sensitivity to distance from grassland across seasons. Together, these findings suggest that maintaining appropriate grassland structure and spatial continuity, with regulated anthropogenic use outside the breeding season, is likely to be an important aspect for sustaining florican populations in human-dominated landscapes.

4.3 Movement strategies in a diffuse migration system

Our results reveal that Lesser floricans employ multiple movement strategies within a broadly diffuse migration system. The contrast between resident and migratory individuals during the non-breeding season adds an additional layer of complexity. The tracked migrants used grasslands less and alternative open habitats more during the non-breeding season than residents, suggesting a broader, more flexible habitat-use strategy. Residents, in contrast, remained more tightly linked to grasslands year-round. This divergence suggests that individuals adopt different solutions to the same fragmented landscape: some remain closely associated with core grassland areas, while others, mostly migrants, range more widely and exploit a broader suite of open habitats. This might be a trade-off between preferred habitat and survival strategies, linked to the non-availability or limited presence of grassland habitats in the migration areas. Moreover, even tracked migrants showed very limited use of closed or heavily modified habitats, underscoring the fundamental importance of open landscapes for the species.

The two migratory males tracked over multiple years (LFM9 and LFM10) followed contrasting movement strategies: LFM10 exhibited a relatively structured, repeatable seasonal strategy between western and southern India, whereas LFM9 showed a more flexible pattern, including a prolonged nomadic phase in central India and a full monsoon stay in Gujarat. Such individual-level variation suggests that migratory behaviour in Lesser floricans is not governed by a single fixed route but instead reflects alternative responses to spatially and temporally variable habitat availability in fragmented agro-grassland landscapes.

At a broader scale, these strategies appear to operate within a spatially diffuse migration system. Rather than following a narrowly defined corridor, Lesser floricans migrate along a broad, flexible flyway spanning western and central India. Diffuse migration systems are common in species such as Eurasian curlew (Numenius arquata), kori bustard (Ardeotis kori) and little bustard inhabiting structurally open landscapes, where the absence of strong geographic barriers allows individuals to select among multiple routes (Mmassy et al., 2022; Guo et al., 2023; González del Portillo et al., 2025; Gregory et al., 2025). The landscape characteristics of western and central India likely facilitates this flexibility. The relatively uniform topography, combined with extensive mosaics of grassland and agriculture, creates a permeable landscape that allows individuals to adjust their routes opportunistically (González del Portillo et al., 2025). Consistent with this, our analysis did not identify clear geographic bottlenecks along the migration axis. Corridor width varied gradually across the flyway, and no strong constrictions were detected. Lesser florican appears to migrate through a landscape that provides multiple viable pathways, resulting in a dispersed movement pattern. This dispersed pattern was also reflected in the high Migration Flexibility Index and the prolonged transit of floricans through regions such as Rajasthan and Madhya Pradesh.

Landscape structure, nevertheless, influenced movement permeability. Tree cover emerged as a significant negative predictor of corridor width relative to grassland habitats, consistent with the ecological specialisation of Lesser floricans as grassland-dependent birds. Cropland showed a marginal negative relationship with corridor width, suggesting that agricultural landscapes may exert moderate resistance to movement, even though they are part of the broader movement matrix. Because migration routes are spatially variable, the persistence of migration likely depends on the availability of multiple permeable pathways rather than the protection of a single corridor. Grassland habitats appear particularly important for maintaining migration flexibility. Preservation and restoration of grassland ecosystems across western and central India may therefore play a crucial role in supporting this species’ long-distance movements. To maximise conservation impact, restoration efforts should be prioritised at the key stopover sites (Supplementary Table 6) and core areas of the migratory corridor identified in this study, thereby safeguarding the functional connectivity required for successful seasonal migrations.

From a conservation perspective, this intraspecific variation does not weaken the case for grassland protection. Instead, it highlights the importance of maintaining a network of suitable habitats across the landscape. As grasslands become increasingly fragmented, a greater proportion of the population may be forced into wider-ranging and potentially riskier movement strategies, with uncertain consequences for survival and reproduction (Sastre et al., 2009; Collar et al., 2017; Ram et al., 2022). Preserving and restoring grassland ecosystems across western and central India will therefore be essential not only to support breeding populations but also to maintain the landscape connectivity required for the species’ long-distance movements.

4.4 Night-time light pollution and anthropogenic pressures

Our analyses of nocturnal habitat use further highlight how different movement strategies mediate responses to human disturbance. Resident Lesser floricans strongly avoided areas with higher night-time light intensity, consistent with the idea that artificial light can disrupt behaviour, alter perceived predation risk, or interfere with natural activity rhythms (Russart and Nelson, 2018; Burt et al., 2023). Migrants, in contrast, showed a weakly positive association with light at night, possibly reflecting incidental use of illuminated agricultural or settlement fringes during broader migratory movements (Ram et al., 2023). An alternative explanation may be that tracked migrants may have limited access to dark areas because migration routes and stopover or wintering landscapes are often embedded within human-modified regions where some degree of artificial illumination is prevalent. Under such conditions, exposure to brighter environments may reflect constrained availability rather than preference. While the mechanisms remain uncertain, this divergence reinforces the idea that movement strategy shapes how individuals interact with anthropogenic landscapes. In increasingly developed rural environments, artificial light is often correlated with other pressures such as traffic, infrastructure, and human activity. The observed association is more likely to reflect the broader influence of human-modified landscapes, of which artificial night-time illumination is one component. Because night-time light is correlated with settlements, roads, croplands, and associated infrastructure, our analysis cannot distinguish the effects of artificial light from other anthropogenic pressures. Future studies incorporating explicit measures of infrastructure, road density, settlement proximity, and artificial lighting will be necessary to disentangle these effects. The strong avoidance shown by residents suggests that light pollution could further reduce the functional availability of otherwise suitable habitats near grasslands, effectively shrinking the usable landscape even where some grassland remains. In this context, minimising linear infrastructure associated with artificial lighting (e.g., near grassland patches or fringes) could help reduce disturbance and maintain landscape permeability during migration. In addition to linear infrastructure, light emissions from villages, including street lighting and household illumination, can contribute substantially to the overall night-time light environment surrounding grassland patches. In this context, reducing unnecessary illumination through use-based and sensor-based lighting strategies may help mitigate impacts.

4.5 Seasonal movement dynamics and ecological drivers

Seasonal differences in step lengths and movement speeds provide additional insight into how Lesser floricans navigate this rapidly changing landscape. Movements were longest and fastest during the monsoon, likely reflecting the combined demands of breeding activities, foraging, and shifting resource distributions associated with seasonal vegetation growth (Dutta and Jhala, 2014), as well as the long-distance transit of migratory individuals into Madhya Pradesh and Rajasthan. During the study period, the southwest monsoon exhibited considerable spatial and temporal variability across western and central India, with delayed onset phases and uneven rainfall distribution reported in several regions (IMD (India Meteorological Department), 2023). Such variability may temporarily alter habitat suitability across landscapes, potentially prompting individuals to track favourable grassland conditions outside traditional breeding areas in Gujarat. The contrasting movement strategies observed in this study may be consistent with this pattern. Longer movement patterns during the monsoon may thus have been an artefact of uneven rainfall, with LFM9 and LFM10 exhibiting movement trajectories in Madhya Pradesh and Rajasthan.

Across all seasons, movement was strongly reduced at night, consistent with diurnal activity patterns typical of ground-dwelling grassland birds as seen in other bustards (Ardeotis nigriceps, Ardeotis kori), which exhibit minimal nocturnal activity and concentrate their movements in daylight hours (Mmassy et al., 2022; Khan et al., 2025). In a fragmented landscape, such seasonal and diel constraints on movement may further limit individuals’ ability to compensate for habitat loss by increasing their range.

4.6 Reconciling movement ecology with land-use change

The Lesser florican has undergone a severe population decline (~80%; Dutta et al., 2018), now persists in only a small fraction of its former range (Rahmani, 2012), and is increasingly confined to a few strongholds characterised by relatively intact grassland. At the same time, agricultural expansion and intensification (BNHS, 2022a; Ram et al., 2023), infrastructure development, overhead power lines (Ram et al., 2022), woody plantations in grassland areas, and mismanaged grazing are major challenges across the species’ range, including in key landscapes of western India (Collar et al., 2017).

The Lesser florican movement results from this study align closely with its current, patchy, fragmented range. The strong spatial associations with grassland patches, the extreme sensitivity to distance from grassland, and the large-ranging behaviour in fragmented landscapes all suggest that further loss or degradation of grassland will have disproportionately large impacts on space use and, by extension, on population viability. The contrasting migratory trajectories observed in LFM9 and LFM10 further indicate that fragmentation not only reduces habitat area but also reshapes movement strategies, potentially increasing exposure to risk during long-distance or exploratory movements.

The evidence that small (4.02 km²) but contiguous grassland blocks can support higher densities than vast, fragmented agro-mosaics further strengthens the case for prioritising the protection and consolidation of remaining grasslands. At the same time, the documented use of croplands and other open habitats outside the breeding season indicates that broader management still requires attention. Measures such as maintaining low-intensity land use around key grassland patches, limiting further fragmentation, reducing disturbance, and retaining structurally suitable open areas could help reduce movement costs and risks.

In this context, community involvement may play an important role in conservation in areas where grasslands are embedded within working pastoral and agricultural systems (Kher and Dutta, 2021), as in our study area. Approaches that integrate local grazing rules and seasonal restrictions on intensive land use could help maintain structurally suitable grasslands while supporting livelihoods, as evidenced in other studies (Gray et al., 2007; Dutta et al., 2011; Dutta and Jhala, 2014). Restoration of gauchars through community-based approaches could play a key role in enhancing habitat availability for Lesser floricans. Such efforts would not only support conservation but also provide sustained fodder resources and livelihood security for local pastoralist communities. However, our results, in concurrence with previous studies (Dutta and Jhala, 2014), indicate that such measures can only complement, not replace, the conservation of grasslands themselves.

4.7 Vertical movement patterns and mortality risk

The low-height (24.4 m) flight patterns observed in this study place Lesser floricans at high risk of collision with overhead power lines, located at the same elevation, as observed in other bustard species (Silva et al., 2014; Shaw et al., 2018; Uddin et al., 2021; Silva et al., 2023). This low-altitude flight behaviour substantially increases susceptibility to collision with overhead infrastructure. Power lines, commonly suspended at 10–40 m, lie directly within the core flight envelope observed in this study. Such overlap has been identified as a major driver of mortality in sympatric bustard species (Uddin et al., 2021). The little bustard likewise demonstrates elevated collision risk in proximity to overhead lines (Silva et al., 2014). Effective mitigation strategies, including flight diverters, line marking, underground cabling, and careful placement of infrastructure away from lekking and foraging sites, have been shown to reduce bustard mortality by 50–80% in comparable systems (Bernardino et al., 2018; Silva et al., 2023). Given the accelerating expansion of renewable energy infrastructure, the present findings underscore the need for spatially targeted risk assessments and proactive implementation of such measures to safeguard the Lesser florican population from collision-related mortalities (BNHS, 2022b; Ram et al., 2022).

5 Conclusion

By linking fine-scale movement behaviour with landscape-scale patterns of land-use change and population decline, this study reinforces a central conclusion: the Lesser florican remains fundamentally a grassland specialist whose persistence is significantly constrained by the amount, configuration, and quality of remaining grassland habitat. While some behavioural flexibility exists through seasonal habitat shifts and divergent migratory strategies, this flexibility is unlikely to offset the ongoing trajectory of grassland loss, fragmentation, and agricultural intensification. In practical terms, the long-term conservation of the species will require halting further alteration of grassland, consolidating and protecting remaining grassland blocks, and managing surrounding agro-grassland mosaics in ways that retain suitable vegetation structure and minimise additional pressures. Without such action, the movement patterns documented here suggest that Lesser floricans will become increasingly confined to smaller, more isolated patches, with rising ecological costs and a risk of further population decline.

A limitation of this study is the small number of tracked migratory individuals (two males and one female), reflecting the rarity of migratory behaviour observed during the study. Consequently, comparisons between migrant and resident birds should be interpreted cautiously, as differences may partly reflect individual variation, sex, geographic origin, transmitter deployment duration, or sampling period rather than migratory strategy itself. Furthermore, movement classifications were based only on the period during which transmitters remained operational, and subsequent movements could not be evaluated. Despite these limitations, the study provides the first satellite telemetry-based assessment of migratory movements and habitat use in the Lesser Florican and establishes an important baseline for future studies incorporating larger sample sizes.

6 Conservation implications

6.1 Centrality of grassland habitats

The findings underscore the crucial importance of intact grassland ecosystems in supporting Lesser florican populations across breeding, non-breeding, and migratory phases. Conservation planning may therefore prioritise the protection and ecological restoration of remnant grasslands across both formally protected areas and multi-use landscapes through community involvement.

6.2 Landscape-level habitat protection and restoration

Conservation interventions could extend beyond state-managed forests to include village pastures, community-managed grazing lands, and degraded grasslands within revenue areas wherever possible. Ecological restoration by removing invasive and woody vegetation may facilitate the recovery of suitable breeding and display habitats (Mohan et al., 2015; Dutta et al., 2018).

6.3 Sustainable grazing and community engagement

Given the species’ close association with structurally suitable grassland patches, carefully regulated grazing management may facilitate territory establishment upon arrival in May–June and provide more favourable conditions for breeding and chick rearing through October–November. Accordingly, temporary seasonal regulation or strategic diversion of livestock grazing to alternative fodder plots, particularly from May to October, could help maintain habitat structure in a condition more suitable for the species while balancing conservation and livelihood considerations. Collaborative approaches involving local pastoral communities and villagers may support habitat restoration and management objectives while sustaining traditional livelihoods (Sankaran, 1997; Dutta et al., 2011; Dutta and Jhala, 2014).

6.4 Incentive-based conservation in agro-grassland systems

Incentive-driven, community-based conservation models, wherever possible, may offer promising pathways for integrating agricultural practices with species conservation. For instance, programs analogous to the Cambodia-based “Ibis Rice” initiative—designed to protect Giant Ibis habitats—demonstrate how market-linked incentives can encourage wildlife-compatible farming while enhancing farmer incomes (Gray et al., 2007). Similar efforts may be adapted to encourage farmers to opt for traditional crops and bustard-friendly agricultural practices within the bustard landscape.

6.5 Mitigating infrastructure-related implications

At broader spatial scales, minimising risks associated with linear infrastructure is desired to maintain habitat functionality. Artificial lighting linked to infrastructure development may alter nocturnal behaviour and reduce the effective availability of otherwise suitable habitats (Burt et al., 2023). Power transmission lines remain a significant source of mortality for bustards; thus, strategic regulation of new installations and prioritisation of suitable mitigation measures in critical habitats and movement corridors are suggested, where feasible (Collar et al., 2017; Dutta et al., 2018; Ram et al., 2022).

6.6 Reducing anthropogenic hazards in agricultural landscapes

Adoption of wildlife-friendly alternatives, such as live hedgerows in place of wired fencing or fishing nets, may help reduce the risk of collisions and entanglement for grassland-dependent avifauna (Ram et al., 2023).

6.7 Towards integrated conservation strategies

Collectively, these approaches highlight the importance of integrated, multi-scalar conservation strategies that combine habitat restoration, community-based land stewardship, and ecologically sensitive infrastructure planning to ensure the long-term persistence of Lesser florican populations.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcosc.2026.1911080/full#supplementary-material

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Footnotes

Edited by: Vitaly Grishchenko, Taras Shevchenko National University of Kyiv, Ukraine

Reviewed by: Yiannis G. Zevgolis, University of the Aegean, Greece

Grzegorz Kopij, Wroclaw University of Environmental and Life Sciences, Poland