Ecotourists' visitation rates and coastal ecosystem restoration: evidence from Tanzania

Ecotourism is increasingly promoted as a conservation financing mechanism in developing coastal regions, yet empirical evidence linking visitor activity to ecosystem outcomes remains limited. This study examined the relationship between ecotourism visitation and land cover change across three Tanzanian coastal sites over 11 years (2012-2022). Using a cross-sectional mixed methods design, we analyzed ecotourism records from 353,492 visitors, assessed landcover changes through supervised classification of multitemporal satellite imagery, and conducted semi-structured interviews with 15 key informants. Strong positive correlations emerged between visitor numbers and restoration investment at Pugu Kazimzumbwi Nature Forest Reserve (r = 0.992), Kaole Ruins (r = 0.971), and Mbudya Island Marine Reserve (r = 0.858). However, land cover analysis revealed contrasting outcomes: Pugu Kazimzumbwi experienced a 10% decline in vegetation and a 9% increase in built-up area despite 45% visitor growth; Kaole Ruins showed farmland conversion, while Mbudya Island gained 10% vegetation through mangrove recovery. Therefore, the findings suggest that, higher levels of ecotourism visitations are positively associated with ecosystems restoration efforts through increased conservation financing and management activities. However, the observed relationships are correlational and should not be interpreted as evidence of direct causality. However, unmanaged tourism also contributes to localized ecosystem degradation. It is essential to strengthen site-specific management strategies, enforce carrying capacity limits, and reinvest tourism revenues into restoration efforts to maximize the ecological benefits of ecotourism in coastal Tanzania.

Ecotourism has emerged as a prominent conservation strategy since the 1980s, offering a potential mechanism to balance ecosystem protection and natural resources preservation (Oliver and Andrea, 2018). Ecotourism has gained international recognition for its role in mitigating habitat degradation, supporting biodiversity conservation, and funding ecosystem restoration efforts (Boley and Green, 2016). In Tanzania, ecotourism expansion has been particularly notable, with growth rates of 10-30% annually from the 1990s through the early 2010s, generating substantial revenue while positioning the sector as integral to both economic development and environmental conservation (Mudasir et al., 2020; Tanzania Forest Services Agency, 2024). Tanzania's coastal ecosystems, encompassing coral reefs, mangroves, estuaries, and seagrass beds, provide critical ecological services, including marine biodiversity support, water quality regulation, coastal erosion buffering, carbon sequestration, and fisheries support (Barbier et al., 2011). However, these ecosystems face severe degradation from overfishing, destructive fishing practices, habitat destruction, industrial and domestic pollution, urbanization, and climate change pressures (Muhando and Rumisha, 2008; Boussougou et al., 2018). Ecotourism has been promoted as a restoration tool in these degraded coastal landscapes, offering opportunities to generate conservation revenue while rehabilitating natural habitats and reinstating ecological interactions (Shen et al., 2023). Previous research in Tanzania's coastal regions has shown that ecotourism provides substantial economic benefits to local communities, creating incentives for ecosystem conservation (Lihuluku et al., 2025).

Despite ecotourism's potential benefits, poorly managed tourism activities can paradoxically intensify ecosystem degradation through visitor trampling, breeding site disturbance, waste generation, and pollution (Barros et al., 2015; Nyaupane and Thapa, 2004). This dual nature of ecotourism as both a conservation enabler and potential degradation driver necessitates careful planning and management to ensure net positive ecological outcomes (De Zoysa, 2022). While extensive research has examined ecotourism in Tanzania's coastal areas, comprehensive empirical studies linking visitor patterns to measurable ecosystem-restoration outcomes are notably absent.

This study examined the effects of ecotourism on coastal ecosystem restoration in Tanzania by assessing how visitor numbers influence restoration investment, analyzing the relationship between visitation rates and ecosystem recovery outcomes, and evaluating ecotourism's contribution to rehabilitating degraded coastal ecosystems across three strategically selected sites representing diverse coastal conservation contexts.

Description of the study area

This study was conducted at three coastal sites in Tanzania's Pwani and Dar es Salaam regions: Pugu Kazimzumbwi Nature Forest Reserve (Kisarawe district), Mbudya Island Marine Reserve (Dar es Salaam), and Kaole Ruins (Bagamoyo) as per Figure 1. These sites represent distinct conservation contexts a protected forest reserve, marine protected area, and cultural heritage site, respectively and were selected for their active ecotourism operations and diverse coastal ecosystem types.

Pugu Kazimzumbwi Nature Forest Reserve (6°3′S, 39°4′E; 120-350m elevation) is located approximately 25 km southwest of Dar es Salaam and encompasses coastal forest ecosystems. Mbudya Island Marine Reserve (6°0′S, 39°3′E; 0-20m elevation) lies 25 km north of Dar es Salaam within the Dar es Salaam Marine Reserve System and features coral reefs and seagrass beds which form part of the biologically important Eastern African Coastal Forest ecosystems recognized for its high biodiversity and endemism (Burgess et al., 1993). Kaole Ruins (6°6′S, 38°5′E; 0-15m elevation) is situated 5 km southeast of Bagamoyo and combines archaeological heritage with coastal mangrove ecosystems dominated by Avicennia marina and Rhizophora mucronata.

The study region experiences a tropical coastal climate with average temperatures of 23-25 °C and bimodal rainfall (800-1,200 mm annually) during March-May and November-December. Local communities depend on artisanal fishing, small-scale agriculture, seaweed farming, and ecotourism activities for their livelihood.

Study design

The research, conducted from March to May 2024 in Tanzania's Pwani and Dar-es-Salaam regions, used a cross-sectional design with purposive sampling and randomization. Both qualitative and quantitative methods were used to collect data. Ecotourism sites were chosen for their significance and categorized into natural and cultural types, including marine parks, reserves, coastal forests, and cultural heritage sites. Selection prioritized activity level and resource diversity across coastal Tanzania.

To examine how ecotourism activities contribute to restoring degraded coastal ecosystems, key informant interviews were held with selected community members living within 50 km of the attraction site. This boundary was chosen because proximity to ecotourism sites often leads to increased community participation in conservation, more direct involvement in ecosystem restoration efforts, and greater awareness of ecological changes caused by ecotourism. Community members in this area are more likely to have direct knowledge of both the degradation processes and restoration results, making them essential sources for understanding how ecotourism affects ecosystem recovery.

Interviews with key informants were systematically organized. A purposive sampling methodology was used to select key informants based on their specific roles and expertise. The sample comprised of 15 participants chosen across three designated attraction sites. Participants were recruited from four principal stakeholder categories: district council representatives, village elders, Tanzania Forest Services Agency (TFS) representatives, and tour operators. The determination of the sample size of 15 participants was guided by the principle of data saturation; whereby subsequent interviews were concluded when additional participants did not provide new information relevant to the study objectives. This methodological approach ensured the comprehensive capture of diverse perspectives from each site while maintaining academic rigor.

Data collection

Data collection involved both primary and secondary sources. Secondary data were obtained from TFS and Marine Park visitor statistics, management plans, policy documents, and annual reports spanning 2012-2022. The primary data were collected through key informants' interviews, utilizing a semi-structured questionnaire. A total of 15 key informants were interviewed, with 5 selected from 3 tourist sites to capture diverse perspectives. The key informants comprised multiple stakeholder categories: three tourism officers (one from each site), three tour guides, three tour operators, three community elders, and three additional tourism staff members. Among the interviewees, eight were male, and seven were female. The ecotourism data collected from the key informants spanned 11 years, from 2012 to 2022. This encompassed general information on historical sites, types of ecotourism activities offered, visitation rates, impacts of ecotourism on the natural ecosystem, and measures implemented to mitigate, halt, or restore ecosystem degradation.

On the other hand, satellite imagery was acquired from USGS Explorer for consistent analysis of Land Use Land Cover (LULC) between 2012 and 2022. The images were selected from Path 166/Row 065, with less than 10% cloud cover to minimize atmospheric interference. The downloaded images were then clipped to the Area of Interest (AOI) using a study-area shapefile to reduce processing time and focus solely on the region of interest. This was done in QGIS using the “Clip Raster by Mask Layer” tool. To enhance radiometric comparability between scenes, raw Digital Numbers (DN) were converted to physical quantities. TOA reflectance values were calculated for all reflective bands using the metadata-provided multiplicative and additive rescaling factors (gain and bias) along with the Solar Zenith Angle. Thermal bands (Band 6 for Landsat 5, Band 10 for Landsat 8) were converted to Brightness Temperature (BT) in Kelvin. LULC classification was performed using the Random Forest (RF) machine-learning algorithm via the Dzetsaka plugin in QGIS. Random Forest is a robust ensemble classifier well suited to multi-band satellite data.

Data analysis

Ecotourists' visitation trend

The raw data were cleaned to identify and remove missing values and duplicates. This step is vital to ensure data accuracy and to address any inconsistencies. Following the cleaning, the data were systematically reorganized and encoded into a format suitable for analysis using SPSS. For instance, categorical variables such as' time in years were encoded numerically from 1 to 11, each number representing a specific year from 2012-2022. Additionally, the data were rearranged so that individual observations were structured into rows, while variables occupied the columns. Upon inputting the data into SPSS, variables were meticulously reverified to confirm they were correctly formatted as ordinal, nominal, or scale variables. Subsequently, data analysis was conducted, and generated outputs including total ecotourist visits, mean annual visitations, maximum and minimum number of visits. Additional out puts were land cover changes and details on restoration initiatives. The results were visualized through charts and tables to facilitate interpretation, comparison, and presentation.

Geographical/satellite data land cover change

Training samples representing key land cover classes were manually digitized using the image scenes. These samples were labeled based on visual interpretation, existing LULC maps, and Google Earth references. Classes typically included built-up/urban, vegetation (forest and agriculture), water bodies, and bare land/exposed soil. Separate training samples were used for each year (2012 and 2022). The Dzetsaka plugin was used to load reflectance bands as input layers, assign training samples, build a virtual band set, create a color composite, and execute classification using Random Forest with default parameters (or customized based on trial accuracy). The classified maps were evaluated for their reliability and correctness using an error/confusion matrix, comparing classified outputs with reference points. The accuracy metrics computed includes User's Accuracy (UA): Probability that a pixel classified into a category represents that category on the ground, Producer's Accuracy (PA): Probability that a pixel of a certain ground class is correctly classified, Overall Accuracy (OA): Percentage of correctly classified pixels across all classes and Kappa Coefficient (κ) which Measures agreement between classification and ground truth while accounting for chance agreement. Validation points were derived from high-resolution imagery and, if available, ancillary reference data. Land cover classes identified through this classification process serve as indicators of restoration in this study. Changes in key classes such as vegetative cover (Forest and agricultural lands), water bodies, and bare land over 2012-2022 were analyzed to evaluate the status and extent of ecosystem restoration. For example, an increase in vegetation cover or a reduction in bare soil over time indicates successful restoration efforts, while expansion of built-up areas or loss of water bodies might signal degradation. These restoration indicators, derived objectively from satellite data, were subsequently correlated with ecotourism visitation rates and documented restoration activities to assess how ecotourism supports ecosystem recovery.

Relationship between visitation rates and ecosystem restoration

To examine the relationship between visitation rates and the restoration of the natural ecosystem visitation data spanning 2012 to 2022 were analyzed using the Statistical Package for Social Sciences (SPSS) version 25 to test for any correlations between visitation rates and ecosystems restoration. Satellite images were processed through supervised classification to monitor ecosystem changes over time.

Pearson's correlation coefficient was calculated to assess the relationship between annual visitation and restoration indicators like restoration scale and land cover changes, measuring the linear correlation to evaluate how visitor numbers align with restoration efforts. Because only 11 annual observations were available for each site, the relationships are interpreted as exploratory rather than causal of the observational study design. Pearson correlation was used as an initial assessment of linear association rather than establish causal relationships.

Ecotourists' visitation and landcover change trends

Ecotourists' visitation trends

Over 11 years study period 2012-2022, the three coastal sites recorded 353,492 total visits with substantial inter-site variation: Kaole Ruins averaged 21,347 visitors annually, Mbudya Island 15,856, and Pugu Kazimzumbwi 5,238. The results show that among the three study areas, Kaole Ruins had the highest number of visitors over 11 years, followed by Mbudya and Pugu-Kazimzumbwi. Mbudya Island has a steady ecotourism presence throughout the study period (2012-2022), supporting consistent revenue for marine conservation. The Kaole ruins had no recorded visitors until 2016 due to delayed promotion and underdeveloped infrastructures but experienced a rapid increase from 2017 to 2019 to 2022. Pugu Kazimzumbwi had no record of visitors until 2017, owing to limited access and awareness, with minimal visitation initially and exponential increases in visitation in 2019. The average annual visitation rate across all three regions was 10,711.88, with Kaole Ruins having the highest yearly visitation rate relative to other destination sites. The maximum annual visitation rate was observed at Kaole ruins in 2022. The trends in visitors and activities are presented in Tables 1, 2, respectively. Pugu-Kazimzubwi and Mbudya accounted for the majority of ecotourism activities (10), whereas Kaole ruins had the fewest (4). Out of all the ecotourism activities, worshiping had the highest visitation rate as observed in Kaole Ruins across the 11 years, followed by hiking within Pugu Kazimzumbwi. Most visitors to all destination sites were male (Table 3).

Land cover change trends

The results from land cover changes analysis revealed distinct patterns across all three sites between 2012 and 2022. In Pugu-Kazimzumbwi, vegetation coverage decreased significantly from 49 to 39%, while built-up areas nearly doubled from 11 to 20%, bare land expanded from 25 to 35%, and waterbodies shrank from 15 to 6%. Kaole Ruins experienced the most dramatic change in farmland coverage, which plummeted from 35 to 10%, alongside a substantial increase in built-up areas from 5 to 22%, a slight decrease in vegetation from 35 to 30%, an increase in bare land from 10 to 22%, and nearly doubled flooded vegetation from 5 to 9%, with waterbodies slightly decreasing from 10 to 7%. Whereas, Mbudya Island, despite being predominantly water, showed a reduction in water bodies from 60 to 50%, an increase in vegetation from 25 to 35%, a doubling of rangeland from 5 to 10%, and a halving of bare land from 10 to 5%. The results on land cover change patterns in the three study sites showed an improved classification accuracy across all sites, reaching the highest user accuracy for vegetation in Pugu-Kazimzumbwi (94.9%) and for waterbodies in both Kaole Ruins (92.9%) and Mbudya Island (96.0%) by 2022.The percentage landcover changes for years 2012 and 2022 are presented in Figures 2, 3, and Table 4.

Correlations between the number of visitors and ecosystem recovery

The results revealed a strong positive correlation between the visitation rate and restoration scale across all three sites, indicating that higher visitor numbers are associated with higher restoration scales (Tables 5, 6) suggesting that sites with higher annual visitation generally recorded higher restoration score during the study period. Moreover, the results showed a direct correlation between visitation rates and restoration index across all sites. The results showing positive correlations provide preliminary evidence of an association between visitation and restoration indicators. The relationships are exploratory rather than causal of the observational study design. Pearson correlation was used as an initial assessment of linear association rather than establish causal relationships.

Restoration initiatives in the study area

The result revealed three major restoration initiatives (Table 6). The restoration initiatives have been systematically organized through partnerships by involving local communities, government agencies, conservation organizations, and ecotourism stakeholders. At Pugu Kazimzumbwi and Kaole historical sites, tree planting is organized by TFS, TAWA, LGAs, and WWF.TFS play an important role in providing technical assistance on raising, managing, distributing, and planting the trees. The community members provide labor for nursery establishment, management and tree planting activities. Native tree species are selected and planted to restore degraded forest areas, and tree planting activities are often scheduled to coincide with community events and ecotourism seasons. Projects like Jane Goodall of Pugu Kazimzumbwi provide fund for the establishment and management of tree nurseries. Mbudya Island focuses on marine conservation with regular beach clean-ups, and coral reefs restoration coordinated by local government authorities (LGAs) in collaboration with ecotourism operators and volunteers to maintain the island's pristine coastal ecosystem. The extent of these activities varies by site, with large-scale tree planting and coral restoration efforts at Pugu Kazimzumbwi and Kaole, and focused, periodic beach cleanups at Mbudya Island. Importantly, these restoration efforts are closely linked to ecotourism; significant funding for tree planting is sourced directly from ecotourism revenues, and coral reef restoration funding correlates positively with visitor growth, indicating that ecotourism development incentivizes investment in ecosystem conservation. Beach cleanup programs are also driven by the need to sustain the attractiveness of ecotourism destinations. This substantiates that ecotourism not only benefits local livelihoods but also plays a critical role in supporting and financing ecosystem restoration initiatives at these sites.

The results reveal a positive trend in ecotourism visitation over the 11 years (2012–2022), with varying rates across the three sites. The visitation trends indicate a general increase in annual visitors at all three locations. However, there was a sharp decline in Mbudya's visitation in 2019, followed by a steady rise in subsequent years. This pattern suggests possible external factors, such as regulatory changes, ecosystem events, or socio-economic challenges, which could have disrupted the flow of visitors (El Moslem Badr, 2022). This highlights the need for adaptive management strategies in ecotourism areas to respond effectively to such fluctuations, similar to the challenges faced in other regions of Tanzania, where infrastructure and accessibility issues limit eco-tourism potential (Pasape, 2022).

The visitor's activity trends across three sites revealed complete differentiation with overlap in visitor experiences, as demonstrated by the activity participation data from 2012-2022. The result shows absolute specialization where each site offers unique activities with no shared experiences between destinations. Analysis of visitor activity data shows that each site has developed specialized tourism niches. Three sites offer distinct and non-overlapping ecotourism activities that reflect their unique environmental characteristics. Pugu-Kazimzumbwi provides forest-based activities, including hiking, bird watching camping, kayaking, fresh air and meditation, and sport fishing. These activities are absent at the other two sites, with zero participants recorded for any of these activities at Mbudya Island or Kaole Ruins. Kaole Ruins demonstrates complete cultural and religious tourism specialization, with worshiping activities dominating visitor experiences and heritage tours along with some filming activities. None of these cultural activities occur at the other sites, showing zero overlap in visitor experiences. Mbudya Island maintains exclusive marine-based tourism activities, including snorkeling, beach picnics, diving, boat tours, photographing, and sunbathing. These marine activities are completely absent from both Pugu-Kazimzumbwi and Kaole Ruins. This complete differentiation indicates that the sites complement rather than compete with each other, as they serve different tourism motivations and visitor preferences. The zero values across all non-native activities for each site confirm that there is no overlap in visitor experiences, making each destination irreplaceable within the regional tourism system. Ecotourism activities such as hiking, snorkeling, bird watching, and worshiping are highly popular in these areas, with Kaole Ruins being the least visited among the three study areas. Notably, Kaole Ruins and Mbudya had the highest number of visitors and diversity of activities as Pugu Kazimzumbwi, which aligns with their earlier availability for ecotourism and diverse activities (Pasape, 2022).

Pugu-Kazimzumbwi Nature Forest Reserve operates under centralized government management through the Tanzania Forest Services (TFS) Agency since 2010. The conservation effectiveness observed at Pugu-Kazimzumbwi today can be traced to significant institutional management evolution. Established in 1936, the forest reserve initially operated under colonial forestry regulations that permitted both agriculture and commercial logging activities, with intensive logging continuing until the 1970s (Gwegime et al., 2013). During the post-independence period through the early 2000s, management responsibility fell under the broader Ministry of Natural Resources mandate, which lacked the specialized forest conservation focus and dedicated resources necessary for effective protected area management. This management gap resulted in continued forest degradation, agricultural encroachment, and uncontrolled resource extraction that reduced the original forest cover significantly as reported by Mdemu and Burra (2016). The transition to Tanzania Forest Services (TFS) management marked a critical turning point in conservation outcomes. TFS brought specialized forestry expertise, dedicated conservation protocols, and structured community engagement programs that addressed the root causes of previous degradation. Currently, the site is managed by a District Forest Conservator (DFC) supported by Forest Officers and Forest Guards who conduct regular patrols and boundary maintenance. Decision-making follows a hierarchical structure, with decisions made at the Eastern Zone headed by the Zonal Conservation Commander (ZCC) and TFS headquarters in Dodoma and implemented through TFS Kisarawe District offices. The operational activities are guided by Management plan, Institution's Strategic plan III. The management budget originates from its sources, and others come from key donors such as the United Nations Development Fund for Nature (UNDP) and WWF, primarily allocated for office running, patrol operations, and infrastructure maintenance, including ecotourism infrastructure development. Community engagement occurs through Village Forest Committees established in five adjacent villages, though their role remains largely consultative rather than participatory in management decisions. The management approach emphasizes forest protection through law enforcement. However, community relations remain strained due to limited benefit-sharing mechanisms and restricted access to traditional forest resources.

Kaole Ruins is under the Department of Antiquities within the Ministry of Natural Resources and Tourism. Dating to the thirteenth to sixteenth centuries, the site's historical and cultural significance was always apparent, but early management under the broader Ministry of Education and Culture provided only basic protection. Limited site interpretation, inadequate visitor facilities, minimal security, and lack of specialized archaeological conservation expertise meant that the site's potential for both heritage preservation and educational tourism remained largely unrealized. The transfer to the National Museums of Tanzania management brought the specialized archaeological and heritage conservation expertise necessary for effective site stewardship. Professional curation, comprehensive site conservation protocols, enhanced visitor education programs, and systematic maintenance have transformed Kaole into a model heritage site. Since 2020, operational management has been delegated to TFS, marking the shift from direct management by the Antiquities division. The transition to TFS aimed to strengthen infrastructure development and enhance specialized heritage management compared to the earlier period. The site is managed by a Site Conservator supported by other staff, including the tour guides, curators, and Site Guards. Management decisions are made collaboratively between the local site management and the Director of Antiquities office in Dar es Salaam, TFS Eastern Zone, and headquarters, with input from the Bagamoyo District Council.

Mbudya Island falls under the jurisdiction of the Dar es Salaam Marine Reserve System, with oversight by the Marine Parks and Reserves Unit (MPRU). Prior to marine parks management, the island operated under general fisheries regulations administered by the Ministry of Natural Resources and Tourism's fisheries division. While this provided basic legal protection, it lacked the specialized marine conservation expertise, dedicated enforcement capacity, and integrated ecosystem management approach necessary for comprehensive island and marine habitat protection. Limited resources and competing sectoral priorities meant that conservation activities were often reactive rather than proactive. The establishment of the Marine Parks and Reserves Unit (MPRU) created the institutional framework necessary to realize Mbudya's conservation potential. MPRU's specialized marine conservation mandate, dedicated personnel, and integrated management approach addressing both terrestrial island ecosystems and surrounding marine environments have enabled systematic protection of the island. Regular monitoring programs, organized tourism management, and community engagement initiatives demonstrate how specialized institutional capacity enhances conservation outcomes while maintaining the island's inherent ecological significance. The site employs a comprehensive management structure that includes a Marine Park Warden, Marine Rangers, and a Community Liaison Officer working together. Decision-making processes utilize monthly multi-stakeholder consultations that bring together MPRU staff, local government officials, and community representatives. Financial resources are distributed across marine patrol activities, community development programs, research and monitoring initiatives, and facility maintenance.

Local participation is facilitated through Beach Management Units (BMUs), which engage fishing communities in both conservation efforts and tourism activities. This collaborative management model has generated significant local ownership, with community members serving as marine patrol assistants and tourist guides. These findings have important implications for conservation strategy in Tanzania and similar contexts. The success demonstrated at all three sites indicates that investing in specialized institutional capacity, whether for forest, marine, or heritage conservation, represents a critical pathway to achieving conservation objectives. Rather than seeking to change or enhance the fundamental characteristics that make sites valuable for conservation, management efforts should focus on developing institutional frameworks with the expertise, resources, and mandate necessary to protect existing values effectively. This approach suggests that conservation success is achievable across diverse ecosystems and heritage types when appropriate institutional capacity is developed and sustained over time.

The conservation effectiveness observed across all three sites demonstrates a common pattern: early management under broader institutional mandates (colonial forestry, general tourism, cultural affairs) provided basic protection but lacked specialized conservation focus. The transition to specialized agencies (TFS for forests and Heritage sites, MPRU for marine areas) created the institutional capacity necessary for effective conservation. This institutional evolution, rather than changing the fundamental ecological or cultural value of these areas, enhanced the management systems needed to protect and sustain these values.

The comparative analysis demonstrates that recently developed ecotourism sites show stronger correlations between visitor numbers and ecosystem conservation than locations with longer tourism histories. However, this correlation does not necessarily indicate positive ecosystem outcomes, as evidenced by the accelerated landscape transformation observed particularly in ecologically sensitive areas like Pugu Kazimzumbwi's forests. The stronger correlations may reflect the absence of established management frameworks and carrying capacity controls that typically develop over time in mature tourism destinations Leung et al. (2018). While all three sites have basic management structures in place, the newer tourism sites appear to lack the refined management protocols and visitor impact mitigation strategies that have evolved at more established destinations (Leung et al., 2018). This suggests that rapid ecotourism growth in recently developed sites can lead to accelerated ecosystem change when management systems have not yet adapted to handle increasing visitor pressures. The threshold effect observed indicates that without proactive management intervention, newly popular ecotourism destinations may experience disproportionate ecosystem impacts during their initial development phase. These findings emphasize the need for site-specific management strategies that consider both the trajectory and pace of tourism development at each location (Tourism Visitation and Land Cover Change Analysis, 2022). As emphasized by Wilson et al. (2024) in their comprehensive review of tourism carrying capacity frameworks, adaptive management approaches that account for site-specific development histories are essential for sustainable tourism planning in ecosystem-sensitive destinations.

Out of all the ecotourism activities, worship emerged as the most popular activity, particularly at Kaole Ruins, with a total of 2,118 visitors participating over the study period. This suggests that spiritual ecotourism forms a vital component of the ecotourism sector in this region. Hiking also had considerable visitation, especially in Pugu-Kazimzubwi, reflecting the appeal of nature-based activities that allow visitors to experience the landscape up close. While some activities like kayaking, bird watching, and camping recorded lower visitation rates, they are integral to diversifying the ecotourism experience and catering to different visitor interests. This variety in activities helps to distribute visitor pressure across the sites and reduce the likelihood of overuse of specific areas, aligning with principles of sustainable tourism that emphasize the importance of diverse and responsible travel practices (Nelson, 2004). While both male and female visitors participated in ecotourism activities at all destinations, male visitors constituted the larger proportion of visitors overall. Therefore, the findings suggest that ecotourism in the study area attracts visitors of both genders but with a male predominance, rather than an equal or nearly equal gender distribution. This indicates a broad appeal of ecotourism activities in the region to all genders, showcasing the inclusivity of the ecotourism experiences offered (Rajasenan et al., 2012).

The correlation analysis indicates positive associations between visitation and restoration indicators and reveals compelling relationships between ecosystem restoration efforts and visitation patterns across the three studied sites. For example, the sharp rise in visitors at Pugu-Kazimzubwi in 2022 correlates with enhanced restoration efforts, such as tree planting and habitat protection. At Mbudya Island, the continuous visitation growth created economic incentives that enabled funding for marine conservation initiatives, including coral reef protection and biodiversity conservation (Nelson, 2004; Mboghoina and Simba, 2024). Strong positive correlations were observed at all three sites, which aligns with previous research suggesting that ecosystem quality significantly influences visitor attraction (Doran et al., 2015). As one key informant from Pugu-Kazimzumbwi noted, “When we started the reforestation program, more visitors came to see the young trees and participate in planting activities,” while a tour guide from Mbudya Island observed that “tourists are more willing to visit when they can see healthy coral reefs during snorkeling.” Area visitation often increases following visible restoration efforts, creating what is termed as a “virtuous cycle” of conservation and ecotourism. The exceptionally strong correlation at Pugu Kazimzumbwi suggests that this forest reserve may represent a typical case of successful integration between ecological restoration and sustainable tourism. This aligns with (Buckley, 2012) framework of conservation tourism, where visitor experiences are enhanced by visible evidence of ecosystem recovery. Additionally, notable gender disparities in visitation responses were observed, particularly at Mbudya Island, where male visitation strongly correlates with restoration, while female visitation does not. This pattern requires careful interpretation, considering the nature of activities at each site and safety considerations. Mbudya Island's marine-based activities, such as diving, snorkeling, and boat tours, may be perceived as more physically demanding or potentially risky, which could influence female participation rates differently than restoration visibility. Field observations revealed that male visitors were more likely to engage in water-based activities that directly showcase marine conservation efforts, while female visitors often preferred beach-based activities that may be less directly connected to visible restoration outcomes. A female respondent noted, “I prefer staying on the beach where I feel safer, rather than going into deep water to see the coral restoration.” This pattern differs markedly from both Pugu Kazimzumbwi and Kaole Ruins, where all genders show similar strong correlations with restoration efforts, likely due to the land-based nature of activities and perceived safety of hiking trails and cultural site visits. This finding mirrors what Doran et al. (2015) identified as gender-specific preferences in ecotourism, where activity type and safety perceptions significantly influence participation patterns.

The results reveal complex dynamics between ecotourism development, land cover change, and restoration initiatives across the three sites, each demonstrating distinct development trajectories and ecological responses. This multi-faceted relationship offers valuable insights for sustainable ecotourism management in ecosystem sensitive areas (Pedroso and Biu Kung'u, 2019). The analysis reveals significant land cover transformations across all three study sites between 2012-2022, demonstrating clear urbanization trends and human-induced ecosystem changes. These land cover transformation are consistent with broader understanding that changes in land use and land cover are shaped by interactions between biophysical and human-related factors (Bajocco et al., 2016). The substantial increase in built-up areas in Pugu-Kazimzumbwi (11% to 20%) and Kaole Ruins (5% to 22%) reflects rapid urbanization in coastal Tanzania, occurring primarily at the expense of natural vegetation and agricultural land (Gwegime et al., 2013). The concerning 10% reduction in vegetation cover in Pugu-Kazimzumbwi Forest Reserve, similar trend of degradation and land cover conversion in Pugu Kazimzumbwi has previously been reported by Mdemu and Burra (2016), indicating that despite ongoing conservation and restoration efforts, development pressures may be outpacing conservation activities in certain areas of the reserve. While restoration activities are occurring, they may be insufficient to counteract broader deforestation pressures or may be concentrated in specific zones while other areas experience continued degradation. Aslo similar patterns of forest loss and recovery dynamics have been docummented in East African protected areas, highlighting the influence of multiple interracting drivers on conservation outcomes. The dramatic reduction of farmland in Kaole Ruins from 35 to just 10% indicates agricultural lands being converted to tourism-related development and residential areas, reflecting the site's transition from a rural to a tourism-oriented landscape. This transformation is evidenced by the establishment of accommodation facilities such as Kaole homestay, just built within 4.1 miles from Kaole Ruins. Mbudya Island's changes, with decreasing waterbodies (60 to 50%) and increasing vegetation (25 to 35%), suggest complex coastal processes including sedimentation, tidal pattern changes, and managed vegetation restoration efforts as part of marine conservation (Mboghoina and Simba, 2024). The decrease in waterbodies across all sites warrants attention, potentially indicating water resource challenges or climate change impacts on hydrological systems. The improved classification accuracy from 2012 to 2022 (with Kappa coefficients rising from 0.85-0.885 to 0.91-0.93) provides increased confidence in the observed changes. These findings underscore the need for integrated land use planning and conservation strategies to balance development needs with ecosystem preservation in these ecologically and culturally important sites (DeFries et al., 2007).

Community perceptions further contextualize the relationship between ecotourism development and ecosystem outcomes. The generally positive view of ecotourism among residents reflects recognition of its potential to improve funding for conservation while raising awareness about ecosystem issues. These findings collectively emphasize the need for tailored management strategies that consider both the trajectory and pace of ecotourism development at each location. As emphasized by Wilson et al. (2024) in their review of carrying capacity frameworks, adaptive management approaches that account for site-specific development histories are essential for sustainable tourism planning in ecosystem-sensitive destinations.

The strong relationships between tourism development, land cover change, and restoration efforts observed in this study suggest that integrating ecosystem monitoring and restoration planning into tourism development from the earliest stages could help mitigate negative impacts while maximizing conservation benefits. However, the varying intensities of land cover change across the three sites indicate that additional site-specific management interventions may be necessary, particularly in recently developed destinations experiencing rapid visitor growth. The restoration initiatives currently underway, including tree planting at Pugu Kazimzumbwi and coral reef restoration at Mbudya Island, demonstrate promising models for how ecotourism can potentially contribute to, rather than simply extract from, the ecological systems upon which it depends.

This study reveals that ecotourism visitation is positively associated with ecosystem restoration efforts across the study sites of coastal Tanzania through multiple, interconnected mechanisms. Across Pugu Kazimzumbwi, Mbudya Island, and Kaole Ruins, increased visitor numbers generate higher revenues from entrance fees, accommodation, and related tourism services, which are directly reinvested into restoration activities such as tree planting, coral reef rehabilitation, and habitat protection. The strong positive correlations observed across all three sites (r > 0.95) indicate that as visitation increases, both communities and management authorities are more likely to allocate resources toward conservation, creating a self-reinforcing cycle where ecosystem improvements attract further visitation.

The influence of visitation operates via several pathways. Economic incentives drive restoration investment, as seen in the alignment of peak visitation years with increased restoration funding, particularly at Pugu Kazimzumbwi. Visitor demand for high-quality natural experiences also motivates site managers to maintain and enhance ecosystem standards, while increased visitation fosters greater local awareness and participation in conservation. Key informant interviews further confirm that higher visitor numbers directly translate into more resources for restoration efforts.

However, the relationship is not uniformly positive. While restoration activities increase with visitation, the study also documents ecosystem pressures, such as the 10% vegetation loss at Pugu Kazimzumbwi despite ongoing restoration. This dual mechanism underscores that while visitation provides resources for targeted restoration, it can also contribute to degradation if not managed sustainably. The net benefit thus depends on whether restoration efforts can outpace or mitigate the negative impacts associated with increased tourism.

The temporal analysis suggests that the influence of visitation on restoration varies by site maturity and management capacity. Newer sites like Kaole Ruins show stronger correlations between visitation and restoration, while more established destinations like Mbudya Island exhibit stable, systematic funding mechanisms less dependent on year-to-year fluctuations in visitor numbers.

Despite these clear patterns, it is important to acknowledge that this study's cross-sectional and correlational design does not definitively establish causality. While the evidence strongly supports a positive association and plausible causal pathways, future longitudinal and experimental research is needed to clarify the direction and strength of these effects over time. Ecotourism in coastal Tanzania demonstrates significant potential to drive and sustain ecosystem restoration when managed effectively. Continued investment in benefit-sharing mechanisms, ecosystem education, and adaptive management will be essential to maximize positive outcomes and address ongoing challenges. Future research should focus on time-series analysis and mixed-methods approaches to deepen understanding of the causal links between visitation and ecosystem health, ensuring that ecotourism remains a force for both conservation and community development.