Management Effectiveness of Protected Areas in Mitigating Human Disturbance: A Case Study of the Qilian Mountains for 2000–2022

Abstract

Evaluating the management effectiveness of protected areas (PAs) is critical for refining conservation strategies. One of the key components in the management of PA is the regulation of human disturbance. We evaluated the management effectiveness of the Qilian Mountain National Nature Reserve (QMNNR) in mitigating human disturbance for 2000–2022. Human footprint was used as a key indicator of human disturbance. It integrates eight human disturbance factors: built environments, population density, night-time lights, cropland, pastureland, roads, railways, and navigable waterways. Evaluations are conducted across dual spatial dimensions: (1) constructing an equal-area external buffer zone to compare human footprint dynamics inside versus outside the reserve; and (2) testing the hypothesis that “stricter zonation correlates with improved control of human disturbance” by analyzing management gradients across four functional zones (core, buffer, experimental, and peripheral protection zones). Key findings include the following: (1) The increase in human footprint within the reserve was markedly lower than in surrounding areas, with the internal–external human footprint disparity expanding from 2000 to 2022. (2) Spatial analysis reveals concentrated disturbance hotspots in northern buffer zones, whereas only marginal increases occurred in Sunan County within the reserve. (3) Human footprint growth across functional zones followed a clear ascending order: core zone < buffer zone < experimental zone < peripheral protection zone, validating the efficacy of zoned management. Collectively, these results demonstrate that the QMNNR has effectively curbed human disturbance expansion—particularly in its core area—though vigilance is warranted against emerging “ecological island” risks in the northern peripheral zone. The proposed dual-dimensional human footprint assessment framework further offers a standardized evaluation methodology for large-scale PA in mitigating human disturbance.

1. Introduction

Protected areas (PAs) are pivotal to conserving biodiversity, sustaining ecosystem services, and advancing global sustainable development [1,2,3]. According to the world database on PAs, over 250,000 designated PAs now exist worldwide, collectively covering more than 17% of land and 8% of marine areas—a testament to their growing global importance [4,5,6]. Among these, China has established one of the world’s largest national PA networks, with over 10,000 PAs spanning diverse ecosystems [7,8]. Against this backdrop, a critical question emerges: what is the management status of these PAs? Assessing the management effectiveness of these PAs has thus become a priority for international conservation organizations and researchers alike [2,9,10]. For example, the International Union for Conservation of Nature (IUCN) World Commission on PAs has developed frameworks centered on management processes, outlining key elements for assessing PA performance [11]. Inspired by such frameworks, scholars globally have conducted extensive studies on the management effectiveness of PA. Yet, gaps remain: most existing assessments focus on ecological outcomes inside and outside PAs, including changes in population density [12,13], species diversity [1,14], habitat quality [3,15,16], and ecosystem service provision [7,17]. While these existing indicators effectively capture ecological outcomes, such as biodiversity recovery and ecosystem service enhancement, they fall short of comprehensively evaluating the management effectiveness of PA. For instance, observed improvements in biodiversity may stem from climate change or natural succession rather than human interventions. Furthermore, long-term monitoring of species richness or population dynamics demands substantial resources (e.g., specialized equipment, technical personnel, and funding) [18], compounded by inconsistent data standards across regions and time periods that hinder cross-regional and longitudinal comparisons. Consequently, there is an urgent need—as emphasized by organizations like IUCN and national governments—to develop robust management evaluation frameworks, particularly for large-scale PAs, to address the current overreliance on ecological outcomes at the expense of management process assessment.

PAs prioritize biodiversity and ecosystem services primarily through regulating human activities. For example, legal instruments such as China’s Regulations on Nature Reserves [19] explicitly prohibit logging, grazing, mining, tourism, and infrastructure development in core and buffer zones of nature reserves, while restricting polluting facilities in experimental zones [20,21]. Given this global reliance on human activity control to achieve conservation targets [22], quantifying spatiotemporal changes in human disturbance within PAs offers a pragmatic and scalable method to evaluate management outcomes [2,23,24]. Early applications of this approach have been explored by researchers worldwide. Geldmann et al. [25,26] assessed PA effectiveness in mitigating human population density and changes in land use at global scales. Leveraging the global human footprint model [27,28,29], Jones et al. [30,31] analyzed spatiotemporal shifts in multiple human disturbances—population density, land use, road networks, night-time lights, and shipping—within global PAs. They found that over half of the world’s ~250,000 PAs exhibited rising human disturbance inside their boundaries, threatening ecosystem integrity and authenticity. Similarly, Anderson and Mammides [32,33] concluded that global PAs fall short in controlling human pressures and preserving terrestrial wilderness. Expanding on human footprint datasets, Wu et al. [34,35], Liu et al. [36,37], and Li et al. [38,39] evaluated PA performance in managing human activities across 680 PAs in China, 336 in the Yangtze River Economic Belt, and 23 in the Tibet Autonomous Region. Some similar studies are also available in sub-Saharan Africa [23], mangrove protected area [40], and the Caucasus Mountains [2].

These studies provide a strong foundation for evaluating PA management effectiveness in curbing human impacts, particularly offering a straightforward, scalable approach for rapid assessments of large-scale PAs or multiple PAs within a region [41,42]. However, data availability constraints have limited their capacity to account for dominant human disturbance drivers or treat these drivers as static [25,28,29,30]. This, in turn, undermines the reliability and timeliness of their results. With recent advances in remote sensing, big data, artificial intelligence, and machine learning, incorporating more granular human disturbance factors and addressing these gaps have become increasingly feasible [43].

The Qilian Mountain National Nature Reserve (QMNNR)—a flagship PA on the Tibetan Plateau—supports exceptional biodiversity, including endemic species and critical ecological processes that regulate regional water and climate systems [44,45]. Given its vast scale, existing research on the QMNNR—a reserve comprising multiple, spatially disjointed sub-zones—has focused on localized biodiversity changes in select sub-zones [46] or ecosystem service assessments (e.g., soil conservation) [47]. While these studies offer incremental value, these assessments do not fully reflect whether management efforts are adequately implemented, failing to capture the full and long-term management trajectory of the entire reserve, leaving a critical gap in our understanding of its overall efficacy.

Recent advances in data and technology, however, have created an opportunity to address this limitation by evaluating the management effectiveness of QMNNR in mitigating human disturbance. High-resolution remote sensing imagery, spatially gridded population datasets [48], and refined human footprint frameworks [29,49,50,51]—already validated for global PA assessments—now enable systematic, large-scale measurement of human disturbance [52,53]. For QMNNR, these tools allow us to (1) track spatiotemporal changes in human activities across the entire reserve, (2) compare human disturbance dynamics inside versus outside its boundaries, and (3) analyze variations across its four functional zones, i.e., core, buffer, experimental, and peripheral protection zones. Note that these four functional zones are primarily delineated based on the importance of biodiversity and ecosystem services, and therefore do not align perfectly with administrative boundaries. This approach makes it feasible to conduct a comprehensive evaluation of QMNNR in reducing anthropogenic pressures—this objective was previously unattainable with traditional methods.

Against this backdrop, we leverage the latest global human footprint dataset—revised and calibrated to QMNNR’s anthropogenic context—to quantify human disturbance trends. Specifically, we (1) contrast human disturbance levels pre- and post-reserve establishment, (2) compare inside vs. outside the reserve, and (3) evaluate differences across functional zones. By synthesizing these analyses, we aim to objectively assess QMNNR’s effectiveness in mitigating human disturbance. These findings will provide actionable guidance for refining China’s PA policies—including optimizing zonation and management practices—and contribute to global efforts to strengthen PA systems in the face of accelerating biodiversity loss.

2. Background of the Study Area

The QMNNR, perched on the Tibetan Plateau—the “Roof of the World”—straddles the confluence of three major Asian plateaus: the Qinghai–Tibet, Mongolian-Xinjiang, and Loess [54]. As a globally significant forest ecosystem-type reserve, it plays a critical role in regional ecological security: its dense forests act as a vital water tower, regulating the headwaters of the Yellow River and supplying freshwater to over 100 million people downstream [55]. Additionally, QMNNR harbors high biodiversity, serving as a stronghold for endangered species like the snow leopard (Panthera uncia) and the vulnerable Qilian juniper (Juniperus przewalskii) [56].

Spanning Qinghai and Gansu Provinces, QMNNR covers a total area of 345,530 hectares and spans 96°46′–103°45′ east longitude and 36°29′–39°43′ north latitude (Figure 1). Its landscape is dominated by forests, followed by grasslands/pastures, croplands, and wetlands/water bodies—this mosaic of ecosystems collectively provide carbon sequestration, soil retention, and habitat connectivity [54,57].

Figure 1. Location of the study area and four functional zones of the Qilian Mountain Reserve, China. This nature reserve includes multiple, spatially disjointed sub-zones.

Established in 1988, QMNNR was initially designed to curb unregulated resource extraction and protect its fragile alpine ecosystems. Today, it remains a cornerstone of China’s PA network, yet faces escalating human pressures: rising tourism, expanding infrastructure, and climate-driven changes in land use [45,58]. These challenges underscore the urgency of evaluating its effectiveness in mitigating anthropogenic disturbances.

3. Materials and Methods

3.1. Human Footprint Dataset and Its Refinement

The ecosystems of PAs are subject to various human disturbances. These human disturbances can be categorized into key factors including population density [59], land use changes [60], and linear infrastructure [61]. Population density serves as a critical indicator of anthropogenic intensity, exhibiting a strong positive correlation with ecological disturbance levels—higher-density regions typically experience more intense production and livelihood-related impacts, positioning it as a dominant anthropogenic factor [62]. Land use changes, including urbanization, cropping [60], and grazing [63], manifested as shifts in land cover types and represent the most direct signals of human influence on terrestrial ecosystems, with construction land expansion and cropland/pasture conversion identified as primary disturbance sources [64]. Linear infrastructures such as highways and railways exert significant corridor effects, inducing ecological alterations within 5 km buffer zones [61,65]. These factors collectively establish land use and infrastructure as core anthropogenic drivers.

Sanderson et al. [28] proposed a human footprint model that quantifies and assigns values to human disturbance factors (e.g., population density, land use changes, and linear infrastructure) impacting ecosystems. By overlaying these factors, they generated the first global human footprint map. Subsequently, the team refined the model, expanding the static footprint map into a dynamic temporal series covering 1993, 2000, 2005, 2009, 2010, and 2013 [27,29]. This dynamic framework and dataset has been widely applied in biodiversity conservation and ecological engineering effectiveness assessment [38,66], providing critical insights into spatiotemporal patterns of human–environment interactions.

Despite its widespread application, validation studies reveal spatial accuracy heterogeneity in human footprint models, particularly in ecologically fragile regions like the Tibetan Plateau, where complex terrain and sparse monitoring networks amplify errors [50,67,68]. To address these limitations, Mu et al. [69] introduced a spatiotemporal enhancement framework building upon Sanderson and Venter’s methodologies [27,28,29], incorporating updated data sources and improving temporal resolution to annual updates. This framework generates a global human footprint dataset spanning 2000–2022 with unparalleled temporal coverage and spatial precision (1 km resolution), outperforming existing datasets in detecting subtle anthropogenic disturbances. The dataset quantifies eight human activity factors (Table 1), enhancing Sanderson et al.’s [28] human footprint model, where minimum values, 0, indicate minimal disturbance and maximum values, 50, reflect the greatest human disturbance. For more methodological details on factor quantification and integration, refer to Mu et al. [69].

Table 1. The framework of mapping annual Human Footprint for 2000–2022. Revised from Mu et al. [69].

Human Pressure	Score	Details	Temporal Range
Built environments	0, 4, 10	The pressure score for pixels with urban fractions above 20% was assigned as 10; otherwise, it was assigned as 4.	2000–2022
Population density	0–10	$\left\{\begin{array}{l}10,population\left(P\right)\ge 1000\\ 3.333\times \log \left(P+1\right),0<P<1000\end{array}\right.$	2000–2022
Night-time lights	0–10	Assigned from 0 to 10 according to intervals determined by ten equal quantiles	2000–2022
Croplands	0, 7	The pressure score for pixels with crop fraction above 20% was assigned as 7; otherwise, it was assigned as 4.	2000–2022
Pastures	0, 4	Fraction of pasture in each grid multiplied by 4	2000
Roads	0–8	$\left\{\begin{array}{l}8,distance\left(D\right)\le 0.5\\ 3.75\times \exp \left(-1\times \left(D-1\right)\right)+0.25,0.5<D<15\end{array}\right.$	2000–2022
Railways	0, 8	$\left\{\begin{array}{l}0,distance\left(D\right)>0.5\\ 8,D\le 0.5\end{array}\right.$	2000–2022
Navigable waterways	0–4	$\left\{\begin{array}{l}0,distance\left(D\right)>15\\ 4\times \exp \left(-1\times D\right),D<15\end{array}\right.$	2000–2022

While the dataset exhibits improvements in precision and temporal resolution, it assumes no changes in grazing activities for 2000–2022 due to unavailable pasture data (Table 1). However, grazing constitutes a critical anthropogenic disturbance in the Qilian Mountains. It is a part of the pastoral area on the Qinghai–Tibet Plateau, home to a large number of herders who raise vast amounts of livestock. To address this limitation, this study incorporates the high-resolution grazing intensity dataset of the Qinghai–Tibet Plateau [63] to characterize grazing patterns. Using the same quantification methodology as population density factor in the human footprint dataset [69], we assign disturbance values to grazing activities and overlay them onto the original human footprint dataset, generating a revised version that explicitly accounts for spatiotemporal variations in grazing pressure. This refinement enhances the dataset’s applicability for ecological impact assessments in regions like the Qilian Mountains, where grazing dynamics significantly influence ecosystem sustainability.

3.2. Evaluation of Management Effectiveness in Mitigating Human Disturbance

This study establishes a dual-validation framework to assess the efficacy of nature reserves in mitigating anthropogenic disturbances. It is specified as follows.

(1) Boundary effect validation. A spatial contrast analysis was conducted by constructing buffer zones equivalent in area to the reserves along their perimeters (Figure 1). The QMNNR is located in the northeastern part of the Tibetan Plateau. It and its surrounding areas are characterized by high altitude, cold temperatures, and arid conditions. We further calculated natural environmental factors—including elevation, slope, mean annual precipitation, and mean annual land surface temperature—and human factors such as distance to roads, and land use/land cover types for both the reserve interior and its equivalent-area peripheral buffer zones. Overall, these natural and human factors in the buffer zones outside the reserve were relatively similar to those inside, with no significant differences. This allowed us to conduct a comparative analysis of human disturbance intensity inside versus outside the reserve, revealing the effectiveness of the reserve’s management in mitigating human interference. This method has also been applied in regions such as southern Ecuador [66] and the Tibetan Plateau of China [67]. The dynamic difference in average human footprint values between buffer zones and reserves was quantified using the following model:

$$\Delta HF\left(t_j\right)={\displaystyle \frac{{\displaystyle \sum _{i_{outer}}HF\left(i_{outer},t_j\right)}}{m}}-{\displaystyle \frac{{\displaystyle \sum _{i_{inner}}HF\left(i_{inner},t_j\right)}}{n}}$$

(1)

where ΔHF(t_j) represents the difference in the human footprint value across reserve boundaries in year t_j, and HF(i_outer, t_j) and HF(i_inner, t_j) denote the human footprint values for outer buffer zones and reserve interiors in year t_j, respectively. The variable t_j denotes the specific years 2000, 2005, 2010, 2015, 2020, and 2022. The variables m and n represent the counts of 1 km pixels within the equal-area buffer zone outside and inside the QMNNR, respectively. The spatial aggregation method employed was the built-in mean algorithm within the ArcGIS software platform (version 10.6), which computes the average value of all pixels within each specified zone. Negligible differences in ΔHF(t) suggest inadequate internal controls relative to external pressures; conversely, a sustained increase indicates effective reserve-mediated containment of human pressure expansion.

(2) Functional zoning gradient validation. Following the Regulations on General Planning for Nature Reserves [21] and the importance of biodiversity and ecosystem services in different areas of the nature reserve, the QMNNR was classified into four hierarchical functional zones (Figure 1): the core zone (869.6 km²), strictly protected with all anthropogenic activities prohibited; the buffer zone (562.5 km²), limited to scientific research and monitoring; the experimental zone (1357.4 km²), permitting low-intensity ecotourism; and the peripheral protection zone (665.8 km²), which balances ecological protection with community development. As can be seen from Figure 1, this nature reserve comprises several protected sub-zones. The shapes of these sub-zones vary significantly, and they do not conform to a strict concentric model where the core zone is innermost and surrounded by the buffer zone. In some sub-zones, the core area is nestled deep within, while in others, it is situated on the outer edge.

The different functional zones face varying degrees of human disturbance and threats. Although the core zone is subject to the most stringent management requirements, it does not necessarily face lesser threats. The level of threat primarily depends on the value of the resources within the zone and the effectiveness of external management. For reserves harboring resources of extremely high economic value, the core zone may instead become a “target of all” and face significant, targeted threats from illegal activities. In such cases, the core zone may actually be under far greater threat than the experimental zone. In terms of management, compared to the core zone, the experimental zone allows for certain human activities. Sometimes, management authorities may concentrate their primary manpower and attention on the experimental zone, which has higher human traffic and activity levels (such as handling tourist disputes and managing infrastructure). This could potentially lead to negligence in monitoring the core zone.

Based on the above information and aligned with the Regulations on Nature Reserves of the People’s Republic of China—which establishes a gradient of decreasing management strictness from the core zone to the peripheral protection zone—this study hypothesizes that human footprint changes will follow a corresponding spatial pattern: minimal change or decrease in the core zone, moderate change in the buffer zone, variable change in the experimental zone, and the greatest increase in the peripheral protection zone. To test this regulatory hypothesis, we quantify trends in the human footprint value from 2000 to 2022 across all four functional zones.

Reserve boundaries and functional zoning data were sourced from the National Catalogue of National Nature Reserves of China [8] and the Adjusted Functional Zoning Maps of Qilian Mountain Nature Reserves (Gansu-Qinghai) [39]. Vectorization and topological validation were performed using the ArcGIS platform to ensure spatial analysis precision.

4. Results

4.1. Human Footprint Disparities Inside and Outside the QMNNR

As illustrated in Figure 2, the human footprint value outside the QMNNR consistently exceeded that inside the reserve throughout 2000–2022, with interior human footprint values averaging only 53.2% of exterior measurements. Temporal trends revealed distinct spatial dynamics: while the interior human footprint value exhibited an initial decline from 2000 to 2005 followed by a gradual increase from 2005 to 2022, resulting in a net rise of 0.09—equivalent to a 1.63% increase over 23 years—the exterior human footprint value demonstrated sustained growth interrupted only by a minor decrease from 2005 to 2010, culminating in an overall increase of 1.08, or 11.27%. The widening disparity between the interior—where human footprint value grew from 3.9317 in 2000 to 4.9174 in 2022—and exterior zones underscores the reserve’s moderating role. Despite gradual increases in interior human footprint values, the absolute and relative growth rates were markedly lower than those of the exterior, which saw a +11.27% change compared to the interior’s +1.63%. These findings indicate that the QMNNR has significantly mitigated the expansion of human disturbance, with human footprint value inside the reserve even showing a downward trajectory during 2000–2010.

Figure 2. Changes and disparities in human footprint inside and outside the Qilian Mountain Reserve from 2000 to 2022.

As illustrated in Figure 3a–f, the spatial pattern of human footprint values within and outside the QMNNR exhibited consistent disparities during 2000–2022, with significantly lower human footprint values observed inside the reserve compared to its exterior. Notably, the northern exterior zone demonstrated markedly elevated human activity intensity, contrasting sharply with the reserve’s interior. Within QMNNR, human activities were predominantly concentrated along two highways traversing the northwest and southeast boundaries, while the presence of Sunan Yugur Autonomous County introduced localized anthropogenic impacts. However, the majority of the reserve’s interior maintained human footprint values below 5, with only minimal regions exceeding 13.

Figure 3. Maps of human footprint inside and outside the Qilian Mountain Reserve from 2000 to 2022.

Spatially, the changes in the human footprint value from 2000 to 2022 revealed that the extent and magnitude of the human footprint value increase outside the QMNNR were significantly greater than those inside (Figure 4). The most pronounced rises occurred primarily in areas such as the urban districts of Jiuquan and Jiayuguan cities in the northwest, the Muli Coal Mine region, the urban area of Minle County in the central part, the urban district of Gulang County in the east, and the northern edge of the reserve’s middle section—all with the human footprint value increases exceeding 8. In contrast, the human footprint value within the reserve remained relatively stable overall, with only minor upticks observed in isolated areas like Sunan Yugur Autonomous County and its surroundings. When viewed in phases, the increase in the human footprint was more significant during the period 2010–2022 than in 2000–2010, and the regions showing an increase were also more numerous. This spatial dichotomy underscores the reserve’s role in mitigating human encroachment, particularly in preserving ecologically sensitive areas from extensive anthropogenic influence.

Figure 4. Maps of human footprint changes inside and outside of the Qilian Mountain Reserve for (a) 2000–2010 and (b) 2010–2022.

4.2. Assessment of Human Footprint Variability Across Functional Zones of the QMNNR

The QMNNR encompasses four functional zones, with spatial disparities in human footprint value trends across these zones illustrated in Figure 5. During 2000–2022 across six assessment years, core zones consistently exhibited the lowest human footprint values, followed by buffer zones and peripheral protection zones, while experimental zones recorded the highest human footprint values. Notably, the human footprint value within core zones averaged only two-thirds of those observed in experimental zones, underscoring the efficacy of stricter conservation regulations in mitigating anthropogenic pressures. This zonal gradient highlights the reserve’s hierarchical management framework, where spatial restrictions correlate inversely with human activity intensity.

Figure 5. Human footprint values of four functional zones in the Qilian Mountain Nature Reserve from 2000 to 2022.

Regarding temporal trends, all four functional zones of the QMNNR exhibited a consistent pattern: a decline in human footprint value from 2000 to 2010, followed by an increase from 2010 to 2022, resulting in a slight net rise overall. Among these, the core zone saw the smallest human footprint value increase—rising by 0.0241 (0.58%)—both in absolute terms and percentage growth, making it the least changed zone across the four. The buffer zone and experimental zone recorded larger upticks: 0.0704 (1.58%) and 0.1099 (1.77%), respectively. The peripheral protection zone had the most substantial increase, with a rise of 0.1150 (2.00%). These data demonstrate that human disturbance in all functional zones have been effectively controlled, and crucially, stricter protection regimes correlate with smaller human footprint value increments. While the experimental zone consistently maintained higher human footprint values than the peripheral protection zone from 2000 to 2022, this pattern may reflect pre-2000 or even pre-reserve establishment conditions rather than post-management trends.

5. Discussion

5.1. The Universality of Human Footprint Methodology in Evaluating PA Effectiveness

This study applied the 2000–2022 1 km resolution human footprint dataset to evaluate the long-term management effectiveness of the QMNNR in mitigating human disturbance. The research content of this paper effectively complements the evaluation of ecological outputs in PAs [55,56,57], enabling a more comprehensive and integrated assessment of PA effectiveness. This is particularly relevant for large PAs that are spatially composed of multiple protected zones. By integrating the Qinghai–Tibet Plateau high-resolution grazing intensity dataset [63], we further revised the global human footprint dataset developed by Mu et al. [69], enhancing the reliability of management effectiveness evaluations.

Compared to previous similar studies on the Qilian Mountains [39,44], this study extends data coverage to 2022, offering greater temporal resolution. The findings align overall: human footprint increases within the reserve (1.63%) were significantly lower than outside (11.27%), validating the efficacy of governance (Figure 2). Notably, these conclusions do not contradict media reports highlighting localized ecological degradation in the Qilian Mountains. As shown in Figure 3, localized increases in human footprint are observed in Sunan Yugur Autonomous County and along northwest/southeast highways. However, the core zone of the reserve exhibited only a 0.58% increase (Figure 4), indicating sporadic rather than systemic degradation. This validates the scientific rationale behind the “strict zoning control” strategy—regulations prohibiting industrial facilities in core zones and restricting tourism development in buffer zones effectively reduced anthropogenic disturbances.

Our findings align with and extend evidence from global PA research that utilizes human footprint methodologies. For example, Geldmann et al. [26] conducted a global-level assessment of PAs, demonstrating that over 50% of PAs experienced rising internal human pressures—a pattern echoed in our results, where human footprint value increases within the QMNNR were significantly lower than exterior zones, yet still evident. Similarly, Jones et al. [30] reported that linear infrastructure and land use changes drive human footprint value increases in PAs worldwide; this resonates with our spatial analysis showing human footprint value hotspots along transportation corridors in the QMNNR’s experimental zone. Such parallels underscore the universality of anthropogenic drivers across diverse PA contexts.

However, our results also reveal contextual nuances when compared to other regions. In tropical PAs, such as those in sub-Saharan Africa, Chiaka et al. [23] found that human footprint value increases were primarily linked to agricultural expansion and deforestation, whereas in the Qilian Mountains—a temperate alpine reserve—mining and infrastructure projects were the dominant drivers. This highlights the importance of weighing region-specific factors in human footprint models. Conversely, studies in South America [70] noted that PA effectiveness is often compromised by weak governance, whereas the QMNNR’s strict zoning policy demonstrates the critical role of regulatory enforcement in curbing disturbances—an insight applicable to PAs in developing regions with similar governance challenges.

The dual-dimensional human footprint assessment framework proposed here offers a standardized toolkit for global PA networks. For instance, Kennedy et al. [50] emphasized the need for “managing the middle”—balancing conservation in moderately disturbed zones; this approach mirrored in our findings that experimental and peripheral zones in the QMNNR require targeted interventions.

5.2. Driving Mechanisms Behind Conservation Effectiveness

The suppression of human disturbance increase within the QMNNR can be attributed to the following key mechanisms (Figure 6). First, policy interventions played a decisive role. From 1990 to 2000, Gansu and Qinghai provinces implemented ecological restoration policies such as “mountain closure for forest regeneration” and “grain-for-green programs” to rehabilitate degraded ecosystems. In 2003, the “Qilian Mountain Ecological Migration Project” relocated 12,500 households from core and buffer zones, directly reducing population density and associated grazing pressures. Second, population control measures indirectly curbed human activities. Family planning policies in counties like Gulang and Tianzhu lowered the resident population base within the reserve, further alleviating anthropogenic pressures. Third, technological substitution mitigated grassland degradation. Sunan County promoted “off-site grazing” combined with stall-feeding practices, achieving a 75% adoption rate of enclosed feeding systems. This reduced direct grazing pressure on vulnerable alpine meadows.

Figure 6. Driving mechanisms behind conservation effectiveness of the QMNNR. Light green represents reduction, and light brown represents increase.

Despite these successes, localized increases in human disturbance persisted within the reserve. Two primary factors contributed to this trend: infrastructure expansion and illegal resource exploitation. Under China’s Western Development Strategy (initiated in 2001) [71,72,73], construction of the Ning-Zhang Highway and Lanxin Railway through experimental zones elevated human footprint values to 1.5 times that of core zones (Figure 3a–f). Unregulated mining and hydropower development, exemplified by the Muli Coal Mine (2010–2020), created spatially concentrated hotspots of disturbance. For example, the slight increase in human footprint in Sunan Yugur Autonomous County is primarily due to the penetration of key infrastructure like the Ning-Zhang Highway and Lanxin Railway, which intensified localized human activities. Additionally, historical small-scale mining in this resource-rich area has contributed to residual anthropogenic pressures, leading to a marginal rise in human footprint despite overall stable conditions within the reserve. Additionally, the more significant increase in human footprint during 2010–2022 compared to 2000–2010, along with the wider spread of affected regions, is primarily driven by the intensified implementation of China’s Western Development Strategy post-2010, which accelerated the completion of infrastructure projects such as highways and railways. Compounding this, a surge in unregulated mining activities and hydropower development within the reserve from 2010 to 2020 generated localized hotspots, collectively driving a more pronounced and widespread increase in human footprint compared to the earlier 2000–2010 period.

Unregulated mining, particularly coal extraction, emerged as a critical driver of human footprint value increase in the QMNNR. This aligns with global studies indicating that mining in PAs often arises from economic incentives and weak enforcement. Geldmann et al. [25,26] noted that in developing regions, mineral resources in PAs face higher exploitation risks due to “high profit margins and low regulatory capacity,” a pattern observed in the Qilian Mountains where remote terrain impeded oversight. Additionally, Anderson and Mammides [32] highlighted that collusion between local enterprises and officials often bypasses environmental impact assessments, exacerbating illegal mining. In the QMNNR, such activities were further fueled by energy demands from nearby cities, creating a “resource frontier” effect [58,72].

Small-scale hydropower projects in experimental zones contributed to human footprint value increases. For example, multiple unlicensed hydropower stations were built along rivers in Sunan County, altering hydrological regimes and increasing human footprint value through road construction and grid infrastructure [44,58]. This resonates with findings from Tapia-Armijos et al. [66] in South American PAs, where hydropower development often violates zoning regulations due to policy loopholes and decentralized governance. In the QMNNR, the lack of cumulative impact assessments for multiple small projects allowed incremental habitat degradation [74]. Studies on tropical PAs [23,75] further confirm that hydropower expansion in biodiversity-rich areas often prioritizes short-term economic gains over long-term conservation, echoing the Qilian Mountains’ challenges.

The persistence of these disturbances reflects a broader governance challenge in balancing conservation and development. As Retief et al. [76] argued, PA effectiveness often hinges on integrating “strict legal frameworks with community engagement and transparent enforcement.” In the QMNNR, despite robust regulations (e.g., prohibitions on mining in core zones), implementation gaps allowed localized violations. This aligns with Chiaka et al. [23], who identified “weak monitoring systems and political prioritization of economic projects” as key drivers of PA degradation in sub-Saharan Africa. Conversely, the decline in human footprint value in core zones demonstrates that where policies were strictly enforced (e.g., ecological migration), disturbances were minimized. Thus, the Qilian Mountains case underscores that targeted enforcement and community-centric policies are critical to mitigating human pressures [39,44].

5.3. Policy Implications and Research Outlook

The study reveals critical insights for enhancing the management of the QMNNR. First, human footprint increases in experimental zones and peripheral protected areas are significantly higher than those in core and buffer zones. To address this, future management should prioritize rigorous environmental impact assessments for construction projects in these zones to mitigate habitat fragmentation and biodiversity loss [74,76,77,78]. Notably, human activity growth in northern peripheral regions such as Jiuquan and Jiayuguan—urbanization hotspots—warrants close attention. These areas risk becoming “ecological islands” if unchecked, as economic pressures from protected zones may shift to adjacent regions. Further studies are needed to assess whether such activity increases threaten ecosystem integrity and propose mitigation strategies, such as ecological corridor planning.

Additionally, this study has limitations requiring further investigation. Although the human footprint dataset employed represents the latest version, it omits certain human disturbance factors such as pollution, noise, and tourism activities. Consequently, it provides a conservative estimate of comprehensive human pressure. Future work should integrate additional variables, such as tourism intensity and infrastructure expansion, to improve accuracy. There is room for improving the spatial resolution of this 1 km dataset. Higher-resolution human footprint data, if available in the future, could enable further insights. Additionally, despite its widespread application, the current human footprint dataset suffers from collinearity among the included factors. For instance, areas with high population density often concurrently exhibit extensive built-up environments, croplands, and pastures. While this does not alter the overall trend of our results, it does introduce a bias toward the overestimation of human pressure in some regions. Future studies should employ mathematical models, such as fuzzy algebraic sum [50], to better integrate these factors and minimize inter-factor collinearity. Second, the current analysis aggregates human footprints without distinguishing individual activity types (e.g., mining, tourism). A granular assessment of each factor’s ecological impact would enable targeted regulatory measures. Finally, the cross-sectional comparison overlooks longitudinal interactions between policies and socioeconomic changes. In a protected area with a strict concentric design (core, buffer, and experimental zones), the core zone’s location implies greater remoteness and inaccessibility. It follows that human activity would naturally increase more slowly in the core than in the experimental zone, regardless of management efforts. This spatial autocorrelation may bias the results of our study. Advanced spatial econometric models, such as propensity score matching or difference-in-differences [79,80,81,82], should be employed to disentangle policy effects from exogenous drivers like urbanization, rural-to-urban migration, and spatial autocorrelation. A more robust method would be to select sampling sites both within and outside the protected area that are comparable in terms of their natural and human conditions, and then compare the human footprint datasets between these paired sites. These methodological refinements will strengthen evidence-based decision-making for sustainable conservation in ecologically sensitive regions.

6. Conclusions

As a critical ecological barrier in western China, the QMNNR plays a pivotal role in safeguarding regional biodiversity and ecosystem services. Scientifically evaluating its management effectiveness is essential for optimizing management strategies. This study employed the 1 km resolution human footprint dataset to quantitatively assess the reserve’s capacity to mitigate human disturbance expansion, integrating spatial comparisons between protected and non-protected zones with functional zoning gradients. Key findings are summarized below. Firstly, the human footprint differential between protected and non-protected zones expanded markedly, demonstrating the reserve’s effectiveness in reducing anthropogenic disturbance on core ecosystems. Secondly, human footprint increments exhibited a strong negative correlation with management strictness across functional zones, validating the hypothesis that “stricter zoning correlates with enhanced human activity control.” Thirdly, the proposed dual-dimensional evaluation framework—combining human footprint analysis with spatial comparisons and functional zoning gradients—complements the evaluation results of the ecological outputs of PAs and collectively assists decision-makers in formulating more holistic strategies. This approach provides an efficient tool for rapid macro-level management effectiveness assessment in large-scale protected areas and national reserve networks.