Successful species conservation and management require a comprehensive understanding of species distribution, abundance, habitat preferences, and movement across wider geographic areas, and over long periods (Wernham et al., 1998; Askins et al., 2007; Greenwood, 2007; Bonney et al., 2009; Craigie et al., 2010; Hochachka et al., 2012). However, classical field surveys to collect this information, while effective (Ong’ondo et al., 2022), can be costly and challenging to conduct, especially over large scales (Bland et al., 2015; McKinley et al., 2017). Furthermore, there is often a lack of long-term data collection in many regions – partly due to financial constraints, recurring permits and licenses, and the need to maintain personnel (Bland et al., 2015; McKinley et al., 2017).

Recent studies have highlighted the values of citizen science across multiple fields such as environmental monitoring, emergency response, and the development of management strategies (Savan et al., 2003; Helsloot and Ruitenberg, 2004; Gouveia and Fonseca, 2008; Bonney et al., 2009; McKinley et al., 2017). In addition, studies have investigated the scalability of these initiatives, from local to global levels (Danielsen et al., 2010; Devictor et al., 2010; Hochachka et al., 2012; De Sherbinin et al., 2021). In Africa, there is a growing interest in using citizen science for research and environmental management, as well as an understanding of barriers, benefits, and challenges (Pocock et al., 2019). While the potential of citizen science remains largely untapped in many taxonomies, there is a growing mass collection of citizen science data on African birds (Underhill et al., 2017; Brooks et al., 2022; Lee et al., 2022; Tende et al., 2024). However, with the exception of South Africa, a large proportion of these data have not been fully utilized. This can primarily be attributed to a lack of expertise in Africa in analyzing the often-unstructured nature of citizen science data, and secondly, because different methodologies have been used - namely pentad methods for generic data collection and point-based approaches for a breeding bird survey (Bystrak, 1981). These methodological differences complicate data analysis and require specialized capacity. Furthermore, there is often a failure to frame research questions that can be answered by the available data.

The Kenya Bird Map (KBM), an example of a citizen science initiative, has been instrumental in providing reliable data on bird distribution and relative abundance across Kenya (https://kenya.birdmap.africa). Established in 2013, and based at the National Museums of Kenya (NMK), KBM follows the South African citizen science protocol (Underhill et al., 2017; Brooks et al., 2022). Through the diverse expertise and perspectives of participants, KBM engages volunteers from diverse backgrounds, including experienced and inexperienced birdwatchers, photographers, and wildlife guides, to gather data. Despite its success in recruiting volunteers and accumulating a large dataset for over a decade, KBM has yet to produce comprehensive scientific results. Its data has yet to be analyzed or published, demonstrating a unique and potential area for advancement.

For the first time, this study has utilized the KBM data to predict the abundance and distribution of bustards, storks, and harriers across Laikipia County, Nairobi National Park, and Masai Mara, Kenya. These species primarily depend on open grassland ecosystems, making them particularly vulnerable to habitat changes. The global decline of grassland birds is a compelling concern, largely attributed to changes in land use/land cover (Muchai et al., 2001, 2002, Zhao et al., 2020; Bardgett et al., 2021; Douglas et al., 2023). The three target regions vary in their management levels. Nairobi National Park is a fully protected area under Kenya’s national government, Masai Mara is partially protected (as a large national reserve under the local government) and surrounded by large privately owned land, and Laikipia is managed privately by landowners (Personal communication). These varying management regimes impact how grassland ecosystems are maintained, providing an opportunity for studying ecological dynamics to guide conservation strategies.

Our study was centered on using the KBM data to predict spatial distribution patterns and abundance of grassland specialist birds (bustards) and grassland opportunistic bird species (storks and harriers). We predicted that the extensive spatial coverage and long-term data collection of KBM would make it a valuable resource for understanding the distribution patterns of these grassland birds across Kenya. Specifically, we anticipated identifying key landscape features that influence the occurrence and distribution of bustards, storks, and harriers.

This study covered three regions within Kenya, Africa: Laikipia County, Nairobi National Park, and Masai Mara ( Figure 1 ). Laikipia County, located approximately 200 km north of Nairobi city hosts many conservancies, wildlife reserves, and ranches dedicated to wildlife, livestock production, sustainable land management, and community development (Bersaglio and Enns, 2024). The Laikipia ecosystem comprises several land cover types that support a higher diversity of wildlife species, including threatened and endangered species (Muriithi, 2016). Human activities in Laikipia include traditional pastoralism, agriculture, conservation, and tourism. The recent doubling of private ranches and conservancies has shaped wildlife conservation and land management practices. Additionally, innovative approaches such as community-led conservation initiatives and sustainable land-use practices have been adopted to balance the needs of wildlife and local communities (Bersaglio and Enns, 2024).

Nairobi National Park (NNP), established in 1946 as Kenya’s first national park, covers approximately 118 km² and is located about 7 km southwest of Nairobi’s central business district (Ong’ondo et al., 2025). The park experiences a semi-arid climate with distinct wet and dry seasons, which supports a diverse range of wildlife, including several bird species (Ogutu et al., 2013; Mwangi et al., 2022). Its land cover includes open grasslands, acacia woodlands, shrubland patches, forests, and riverine habitats, although invasive species like Lantana camara and prickly pear cactus threaten native vegetation (Ong’ondo et al., 2025).

Urbanization has encroached particularly from the north and west, while rural areas to the south and east are dominated by grasslands and settlements. These landscape dynamics contribute to habitat fragmentation and pose challenges to wildlife movement. For a more detailed description of Nairobi National Park’s ecosystem, see Ong’ondo et al. (2025).

Maasai Mara, located approximately 300 km southwestern of Nairobi city, comprises Masai Mara National Reserve and several conservancies. The Masai Mara ecosystem is characterized by a diverse landscape that provides habitat to a lot of wildlife species including numerous bird species. Despite its ecological importance, the Masai Mara ecosystem faces challenges attributed primarily to agricultural expansion and settlements outside the reserve (Thompson et al., 2002; Kariuki et al., 2021). Human activities within the Maasai Mara intersect with traditional pastoralism and conservation, which are practiced by the indigenous Maasai community.

We extracted a total of 50,717 records for 16 species reported across the three study regions from 2013 - 2023; these records spread across the three focus groups were: bustards (~20,497 records), harriers (~7,427 records), and storks (~ 22,792 records) ( Table 1 ). Comparing regions, Masai Mara had the highest number of records of all total species (~ 25,799 records), and Nairobi National Park had the least (~ 5,308 records) ( Table 1 ). The three regions had a cumulative 16 species, and 15 species each. The distribution of species varied across the regions, with some being unique to specific regions. For example, the buff-crested bustard (Lophotis gindiana) was only recorded in Laikipia County. The woolly-necked stork (Ciconia episcopus) was recorded in both Nairobi National Park and Masai Mara, but not in Laikipia County ( Table 1 ).

The highest single species records by region included white-bellied bustard (Eupodotis senegalensis) and Kori bustard (Ardeotis kori) (~ 4,222 and 3,000 records, respectively) in Laikipia County; white-bellied bustard (Eupodotis senegalensis) and marabou stork (Leptoptilos crumenifer) (~ 4,274 and 4,077 records, respectively) in Masai Mara; and Marabou stork (Leptoptilos crumenifer) and yellow-billed stork (Mycteria ibis) (~1,873 and 1,027 records, respectively) in Nairobi National Park ( Table 1 ).

The species with the least record by region were African marsh harrier (Circus ranivorus) and Hartlaub’s bustard (Lissotis hartlaubii) (~ 210 and 242 records, respectively) in Laikipia County; western marsh harrier (Circus aeruginosus) and African marsh harrier (Circus ranivorus) (~ 179 and 203 records, respectively) in Masai Mara; African marsh harrier (Circus ranivorus) and Woolly-necked stork (Ciconia episcopus) (~8 and 13 records, respectively) in Nairobi National Park ( Table 1 ). The three regions had different numbers of pentads; Laikipia County had 20 pentads; Masai Mara, 19 Pentads; and Nairobi National Park 3 Pentads ( Table 2 ). In addition, the classification results from Google Earth Engine ( Figure 2 ) illustrate the distribution of land cover types across the three regions and highlight the spatial variation within each study area.

Our regression results indicated that bustards responded positively to patch density, with a mean estimate of 26.3 (SE = 11.7, P < 0.05), and negatively to woodland and shrubland, with mean estimates of -16.1 (SE = 4.7, P < 0.05) and -4.6 (SE = 1.7, P < 0.05), respectively ( Table 3 ; Figure 3 ). However, there was less evidence of grassland having a negative effect on bustards, as indicated by a mean estimate of -8.8 (SE = 4.5, P = 0.05, Table 3 ; Figure 3 ). Storks indicated positive responses to grassland and woodland with mean estimates of 13.6 (SE = 3.8, P < 0.05) and 12.1 (SE = 4.4, P < 0.05) respectively ( Table 3 ; Figure 4 ). However, there was less evidence of patch density having a negative effect on storks, as indicated by a mean estimate of -24.8 (SE = 12.6, P = 0.05, Table 3 ; Figure 4 ). Harriers indicated a negative response to woodland, with mean estimates of -8.8 (SE = 4.1, P < 0.05, Table 3 ; Figure 5 ).

We used KBM data to analyze the spatial and temporal patterns of bustards, storks, and harriers’ abundance and distribution across Kenya’s landscape. Our analysis demonstrated the value of KBM data in predicting species distribution and provided essential knowledge of how specific land cover types affect these species. The research identified key ecological zones supporting bustards, storks, and harriers and quantified the impact of land cover composition (e.g., woodland, grassland) and configuration (e.g., patch density) on their abundance and spatial distribution.

The unique presence of buff-crested bustard (Lophotis gindiana) only in Laikipia County and woolly-necked stork (Ciconia episcopus) in both Nairobi National Park and Masai Mara suggest distinct ecological variation and differences in species distribution across these regions. Kennedy (2014) and Lewis and Pomeroy (2017) described the buff-crested bustard (Lophotis gindiana) as an inhabitant of arid to semiarid climates with stony land cover types. Laikipia County, with its expansive semi-arid landscapes and rocky outcrops, provides these specific habitat conditions preferred by this species, which likely account for the species’ absence from the more humid and densely vegetated regions such as Nairobi National Park and Masai Mara. In contrast, the woolly-necked stork (Ciconia episcopus), was only observed in Nairobi National Park and Masai Mara, both of which offer critical habitat components such as wetlands and open spaces needed for the species’ survival. The absence of such resources in Laikipia County likely restricts the stork’s occurrence in this region. Studies have shown that storks prefer working landscape with wetlands and open spaces as opposed to natural landscape (Johst et al., 2001; Kamiński et al., 2020; Kittur and Sundar, 2020).

The landscape complementation hypothesis (Turner, 1989; Dunning et al., 1992) suggests that species response is favored in landscapes defined by juxtaposition and interspersion, where diverse land cover types containing key resources are closely associated. We argue that the observed positive response of bustards is influenced by both the composition and configuration of the landscape, particularly the proximity and arrangement of grassland, woodland, and shrubland patches, along with patch density. These features likely provide essential ecological resources such as food, shelter, and suitable nesting sites. However, the effect of land cover composition and configuration varied when examined independently, emphasizing that both the types of land cover and their spatial arrangement contribute to the observed species distributions.

García et al. (2007) found that diverse, mosaic land cover benefits fragmented little bustard populations. Our analysis is consistent with this, demonstrating a positive response of bustards to patch density, which underscores the role of spatial arrangement in facilitating suitable habitat conditions (Turner, 1989; Dunning et al., 1992). Higher patch density reflects a mosaic of land cover types with varied spatial arrangements, which offer diverse foraging and nesting opportunities for bustards. The interspersed and juxtaposed land cover types in these landscapes provide both the ecological resources and spatial structure essential for the species’ survival. Woodland and shrubland exhibited significant negative responses on bustard distribution and abundance, highlighting low-quality covers for bustard requirements. These results are consistent with previous studies emphasizing the preference of bustard species for open, less vegetated landscapes (Wolff et al., 2002; Kennedy, 2014; Mmassy et al., 2019), further highlighting the critical importance of habitat configuration in determining species distributions.

The negative, though not significant, effect of grassland on bustard response suggests that grasslands, while less detrimental than woodlands or shrublands, still influence the species’ habitat use. Extensive grassland management practices, particularly those aimed at maximizing livestock feed production in Laikipia County and Masai Mara, may significantly alter the structural composition, disturbance regimes, and spatial arrangement of grassland ecosystems. Bernath-Plaisted et al. (2023) emphasized that grassland bird declines are driven by persistent threats throughout their annual cycle, including habitat loss, agricultural intensification, woody encroachment, and grazing regimes. Our findings align with this perspective, suggesting that these livestock-driven changes disrupt the availability of essential resources for bustards. For instance, intensive grazing and the conversion of grasslands into pastures may reduce suitable nesting sites, limit access to foraging areas, and diminish protective vegetation cover. Moreover, previous studies indicate that bustards show a marked preference for natural, undisturbed grasslands and tend to avoid managed pastures due to frequent disturbances, including livestock grazing and associated human activities (Johnsgard, 1991; Dutta et al., 2010).

The variability in stork responses to different land cover types highlights the suboptimal conditions of current landscapes for supporting stork populations. Our findings suggest that grasslands and woodlands provide critical resources for storks, including foraging areas, water sources, and nesting sites essential for their survival and reproduction. Storks favor open habitats, which facilitate access to prey such as small fish, amphibians, and insects, resources that are more visible and accessible in these environments (Herremans and Herremans-Tonnoeyr, 1993; Gerkmann et al., 2008; Kronenberg et al., 2017; Orłowski et al., 2019). Open spaces also enhance flight efficiency, a crucial factor for storks that forage and migrate over large distances (Gerkmann et al., 2008).

In contrast, dense shrublands limit visibility and mobility and impedes the storks’ ability to detect and capture prey, which may explain the weak positive association observed with this land cover type. Moreover, the negative relationship between patch density and stork response indicates that fragmented landscapes disrupt foraging efficiency by scattering prey too thinly across the habitat, making it more challenging for storks to locate sufficient food. These findings underscore the importance of preserving critical land cover types and ensuring landscape connectivity to maintain functional habitats for storks.

The negative response of harriers to woodland habitats is likely due to changes in both land cover composition and configuration, which reduce the availability of key prey species such as rats, moles, and voles. Changes in vegetation structure, such as excessive tree thinning or dense shrub growth, reduce prey visibility and movement (Morán-López et al., 2015; Zhang et al., 2016). Additionally, fragmentation of woodland habitats, driven by development and land use changes, disrupts prey populations and limits habitat connectivity. These factors, influenced by both composition and configuration, ultimately reduce the habitat’s suitability for harriers. Bersaglio and Enns (2024) argued that the “violent ecological transformation” in Laikipia has led to a simplified and less heterogeneous landscape. This trend has also been documented in other regions (Ndegwa Mundia and Murayama, 2009; Ogutu et al., 2009; Khalid Kija et al., 2020).

For the first time, our study has recognized the importance of ongoing efforts to incorporate KBM data with complementary ecological datasets to enhance our understanding of bird communities and their responses to environmental changes. Utilizing KBM data’s extensive spatial coverage and long-term monitoring capabilities, researchers can elucidate complex ecological relationships and identify critical conservation priorities across various landscapes. This integration between KBM data and environmental variables signifies immense potential for advancing evidence-based conservation initiatives and safeguarding avian biodiversity locally and globally. Moving forward, continued collaboration between researchers, conservation practitioners, and data custodians is essential to utilize the full capacity of KBM in addressing pressing ecological challenges and promoting sustainable conservation and management of natural resources.