Next Article in Journal
Volumetric X-Band Radar Analysis of Acoustic Precipitation Enhancement: A Stratiform Precipitation Case over the Bayinbuluke Basin
Previous Article in Journal
Monitoring Aerosol Dynamics in the Beijing–Tianjin–Hebei Region: A High-Resolution, All-Day AOD Dataset from 2018 to 2023
Previous Article in Special Issue
The Impact of Climate Change on Changes in the Onset and Termination of Growing Seasons and the Area of Agriculturally Usable Land in Slovakia
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Atmosphere
Volume 17
Issue 2
10.3390/atmos17020169
atmosphere-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Coupling Mechanisms Between Vegetation Phenology and Gross Primary Productivity in Alpine Grasslands on the Southern Slope of the Qilian Mountains

by
Fangyu Wang
1,2,†,
Yi Zhang
1,2,†,
Guangchao Cao
1,2,3,
Meiliang Zhao
1,2,* and
Yinggui Wang
1,2
1
Key Laboratory of Tibetan Plateau Land Surface Processes and Ecological Conservation (Ministry of Education), Qinghai Normal University, Xining 810008, China
2
Qinghai Provincial Key Laboratory of Physical Geography and Environmental Process, College of Geographical Science, Qinghai Normal University, Xining 810008, China
3
Academy of Plateau Science and Sustainability, People’s Government of Qinghai Province and Beijing Normal University, Xining 810008, China
*
Author to whom correspondence should be addressed.
†
These authors contributed equally to this work.
Atmosphere 2026, 17(2), 169; https://doi.org/10.3390/atmos17020169
Submission received: 24 December 2025 / Revised: 27 January 2026 / Accepted: 3 February 2026 / Published: 4 February 2026
(This article belongs to the Special Issue Vegetation and Climate Relationships (3rd Edition))
Browse Figures
Versions Notes

Abstract

Understanding the coupling mechanisms between vegetation phenology and carbon productivity is essential for assessing ecosystem responses to climate change and guiding sustainable grassland management. This study focuses on stable alpine grasslands on the southern slope of the Qilian Mountains from 2001 to 2020, a climatically sensitive but relatively under-investigated transition zone on the northeastern Tibetan Plateau. We utilized MODIS NDVI time-series (MOD13Q1) and the latest PML V2 gross primary productivity (GPP) product at 500 m resolution to quantify changes in the start (SOS), end (EOS), and length (LOS) of the growing season. A pixel-wise linear regression approach was applied to evaluate the sensitivity of GPP to phenological metrics, explicitly characterizing how much GPP changes in response to unit shifts in SOS, EOS and LOS. Compared with previous studies that mainly described large-scale correlations between phenology and GPP or relied on coarser GPP products, this study provides a pixel-level, sensitivity-based assessment of phenology–carbon coupling in alpine grasslands using a long-term, phenology–GPP dataset tailored to the Qilian alpine region. The results revealed trends of earlier SOS, delayed EOS, and extended LOS, accompanied by a gradual increase in GPP. However, phenology–GPP coupling exhibited notable spatial heterogeneity. In mid- and low-altitude areas, extended growing seasons enhanced GPP, whereas high-altitude zones showed limited or even negative responses, likely due to climatic constraints such as cold stress and thermal–moisture mismatches. To better understand these spatial differences, we constructed a three-dimensional phenology–GPP sensitivity space and applied k-means clustering to delineate three ecological functional zones: (1) high carbon sink potential, (2) ecologically fragile regions, and (3) neutral buffers. This sensitivity-based functional zonation moves beyond traditional correlation analyses and provides a process-oriented and spatially explicit framework for ecosystem service assessment, carbon sink enhancement and adaptive land-use strategies in sensitive mountain environments.

Graphical Abstract

1. Introduction

Global climate change, characterized by rising temperatures, altered precipitation regimes, and increased frequency of extreme weather events, is profoundly reshaping the structure and functioning of alpine ecosystems [1]. Alpine grasslands, due to their fragile climate adaptability, have been widely recognized as sensitive “indicators” and “amplifiers” of climate change [2]. Vegetation phenology, as a key process in plant life cycles, plays a central role in regulating carbon, water, and energy fluxes, and responds sensitively to climatic variations [3]. In high-latitude and high-elevation regions, earlier spring phenology and prolonged growing seasons have been shown to significantly enhance ecosystem carbon uptake [4,5]. Moreover, phenological changes interact with hydrothermal conditions to modulate the spatiotemporal dynamics and interannual variability of gross primary productivity (GPP) [6,7]. In alpine regions of the Qinghai–Tibet Plateau, however, spring phenology is not solely regulated by air temperature. Snow cover duration and snowmelt timing exert strong controls on the onset of vegetation growth by regulating early-season soil thermal conditions and water availability. Remote sensing studies have demonstrated that reductions in snow depth and earlier snowmelt can substantially advance the start of the growing season (SOS), particularly at high elevations where persistent snow cover may delay phenological development despite regional warming [8,9].
GPP, as the total carbon fixed by plant photosynthesis, serves as a core variable in terrestrial carbon cycling. Compared with net primary productivity (NPP) and net ecosystem productivity (NEP), GPP more directly reflects vegetation photosynthetic potential and exhibits greater accessibility and representativeness at the remote sensing scale [10,11]. In recent years, extracting phenological parameters such as the start (SOS), end (EOS), and length (LOS) of the growing season from remote sensing time-series indices (e.g., NDVI, EVI), and coupling them with GPP, has become an effective approach for revealing vegetation responses to climate change [12,13]. Studies in the Qilian Mountains indicate that earlier spring phenology significantly promotes summer GPP, but the lag effects and feedback mechanisms of phenology-GPP relationships vary across ecological zones [14,15]. Previous studies across the Qinghai–Tibet Plateau have mainly focused on phenological trends and their climatic drivers or on vegetation productivity dynamics analyzed independently from phenology, often at aggregated spatial scales. Although phenology–carbon coupling research has been extensively conducted in North and Northeast China, systematic investigations remain limited in the southern slope of the Qilian Mountains—an ecologically critical zone on the northeastern margin of the Tibetan Plateau characterized by complex terrain and diverse vegetation types [16,17,18,19].
In recent years, substantial efforts have been devoted to characterizing vegetation dynamics over the Qinghai–Tibet Plateau using long-term MODIS time-series data. For example, Qi et al. analyzed the spatial–temporal variations of vegetation phenology in the Qilian Mountains and highlighted the roles of snow cover and elevation in regulating SOS and LOS [8], while Zhou et al. produced a temporally continuous annual vegetation map for the entire Qinghai–Tibet Plateau from 2000 to 2022, providing a consistent land-cover foundation for long-term ecological studies [20]. In parallel, several studies have explored the relationships between vegetation phenology and ecosystem productivity across the plateau, including phenology–NPP linkages and phenology-based productivity proxies [21,22]. More recently, emerging analyses have begun to quantify phenological sensitivities of alpine grassland GPP using advanced statistical or machine-learning approaches [23]. However, most existing studies primarily emphasize vegetation mapping, phenological trends, or productivity dynamics in isolation; rely on productivity proxies or coarser GPP products; or are conducted at broad regional scales. Consequently, pixel-scale quantitative coupling between satellite-derived phenological metrics and high-quality GPP estimates remains insufficiently explored for alpine grasslands on the southern slope of the Qilian Mountains. Alpine grasslands, as the dominant ecosystem type on the Tibetan Plateau and surrounding regions, are ecologically representative in global high-elevation zones due to their large areal coverage, high carbon sink potential, and strong climate sensitivity [24,25]. On the southern slope of the Qilian Mountains, grasslands—mainly alpine meadows and alpine steppes—cover over 80% of the area and represent a major source of regional carbon flux [26,27]. Compared to forests and croplands, grasslands respond more rapidly to temperature and precipitation changes, and their phenological variability significantly influences the duration and intensity of photosynthesis, thereby shaping interannual fluctuations in GPP [28,29,30]. Furthermore, the relatively homogeneous surface structure and lower anthropogenic disturbance of grasslands make them highly suitable for remote sensing-based phenology and GPP monitoring, enhancing both spatial representativeness and ecological interpretability [31]. Studying phenology–carbon coupling in grasslands not only elucidates key processes in climate-sensitive carbon cycling zones but also supports the identification of grassland-specific adaptation strategies and carbon sink enhancement pathways.
Traditional GPP products such as MOD17 often underestimate productivity in alpine regions, limiting their applicability in carbon budget assessments [32,33]. In contrast, the PML V2 product, which incorporates an improved light use efficiency (LUE) model and is driven by satellite and meteorological inputs, demonstrates superior performance and strong agreement with flux observations in the Tibetan Plateau and Qilian alpine grasslands [24,25,26,27,28,29,30,31,32,33,34,35,36,37]. Moreover, scale mismatches between phenological parameters and GPP data often undermine the accuracy of their coupling analyses. The PML V2 dataset offers better spatial–temporal consistency and ecological zone specificity, providing a robust foundation for phenology–carbon interaction studies in alpine ecosystems [38].
With steep topography and distinct climate gradients, the southern slope of the Qilian Mountains is an ideal region for examining phenological and GPP responses to climate forcing. Recent studies have shown that grassland carbon fluxes in this area are jointly influenced by temperature, precipitation and radiation, while phenology exhibits a differentiated “plateau–valley” pattern [39,40,41]. In this study, the alpine grassland ecosystem of the southern slope of the Qilian Mountains is selected as the research area. Using the Google Earth Engine (GEE) platform, stable grassland areas are identified based on the MCD12Q1 land cover product. Time-series NDVI data are extracted from the MOD13Q1 product, and GPP data are obtained from the PML V2 model. At a spatial resolution of 500 m and an annual temporal scale, vegetation phenological metrics (SOS, EOS and LOS) are derived using TIMESAT 3.3 from smoothed NDVI curves, enabling a consistent spatiotemporal match between phenology and carbon flux. Compared with previous studies on the Qinghai–Tibet Plateau that primarily focused on vegetation mapping, phenological trends, or productivity dynamics in isolation, this study (i) targets climatically sensitive but under-investigated alpine grasslands on the southern slope of the Qilian Mountains, (ii) integrates a 20-year satellite-derived phenology dataset with the advanced PML V2 GPP product at the pixel scale and quantifies pixel-wise sensitivities of GPP to SOS, EOS and LOS, and (iii) constructs a three-dimensional phenology–GPP sensitivity space for k-means clustering to delineate functional zones of carbon sink potential and ecological vulnerability. This research aims to advance the understanding of phenology–carbon coupling mechanisms in alpine regions and to provide scientific evidence and technical support for regional carbon sink assessments and climate adaptation strategies.

2. Materials and Methods

2.1. Study Site

The southern slope of the Qilian Mountains is located at the northeastern edge of the Qinghai–Tibet Plateau (98°08′ E–102°38′ E, 37°03′ N–39°05′ N), covering an area of approximately 24,000 km2 (Figure 1). It serves as a transitional zone between arid and alpine regions, characterized by pronounced topographic variation. The landscape comprises diverse geomorphic units, including high terraces, hills, and intermontane basins, with elevations ranging from 2289 to 4705 m [42]. The region exhibits a typical plateau continental climate, with a mean annual temperature of approximately −5.9 °C and average annual precipitation of about 400 mm, most of which occurs between May and September [43]. Dominated by grassland ecosystems with high vegetation coverage, this area represents a key carbon sink region along the northern margin of the Qinghai–Tibet Plateau. Grasslands in the study area mainly consist of alpine meadow and alpine steppe ecosystems, which are widely distributed across the Qinghai–Tibet Plateau and are highly sensitive to climatic variability. Vegetation is dominated by perennial herbaceous species, primarily from the genera Stipa and Carex. Representative species include Stipa purpurea, Carex parvula, and Carex alatauensis.
Due to substantial heterogeneity in topography and climate, grassland phenological dynamics and productivity patterns exhibit marked spatial variation. Moreover, the ecosystem is highly sensitive to climate change, with vegetation phenology significantly influenced by rising temperatures and variable precipitation, thereby playing a pivotal role in regulating regional carbon cycling. As summarized by Wang et al., alpine grasslands across the Qinghai–Tibet Plateau have experienced pronounced changes in structure and productivity under the combined influences of climate change and adaptive management, highlighting the ecological significance of this region for carbon cycle studies [44]. As a critical component of the Qinghai–Tibet Plateau’s ecological security barrier, the carbon productivity dynamics of grasslands in this region not only reflect the response mechanisms of alpine ecosystems to climate change but also bear important implications for regional carbon budget assessments.

2.2. Research Data

This study integrates multiple sources of remote sensing data to analyze the spatiotemporal coupling between vegetation phenology and GPP in the alpine grasslands of the southern Qilian Mountains. Stable grassland areas, defined as pixels classified as grassland (IGBP class code 10) in all years from 2001 to 2020, were identified using the MCD12Q1 land cover product on the GEE platform. GPP data were obtained from the PML V2 product, while NDVI time-series data were extracted from the MOD13Q1 product at a spatial resolution of 500 m. MODIS MOD13Q1 NDVI data were first quality-screened using the accompanying QA layer to remove low-quality observations (e.g., cloud, snow, and aerosol contamination). The remaining high-quality time series showed good temporal continuity, and no additional temporal gap-filling was applied. Using TIMESAT 3.3 software, the NDVI time series was subjected to curve fitting and smoothing by employing the double logistic function combined with a dynamic threshold method to extract key vegetation phenology parameters (SOS, EOS, and LOS). The phenological outputs were converted to GeoTIFF format using ENVI 5.3, and subsequent spatial analysis and modeling were conducted in ArcGIS 10.2 and R 4.5.0. Detailed data sources and descriptions are listed in Table 1.

2.3. Vegetation Phenology Parameter Extraction

Vegetation phenological parameters are key indicators for characterizing the dynamic changes of the growing season. They are widely used to assess vegetation responses to variations in temperature and moisture, and thus play an important role in phenological studies under the context of global climate change.
By comparing the S-G, A-G and D-L filtering methods in areas such as Northeast China and the urban agglomeration of central Yunnan, it was found that the D-L fitting effect was closest to the original curve and had the best reconstruction of the vegetation growth peak [41]. Therefore, in this study, TIMESAT 3.3 software was employed to process NDVI time series data. A double logistic (D-L) function was applied for curve fitting and smoothing, effectively reducing noise and outliers, thereby improving the accuracy of phenology detection. Among the existing phenology extraction approaches, commonly used techniques include threshold methods, moving average methods, and derivative-based methods [45]. Considering that the south slope of Qilian Mountains has dramatic topographic relief, significant climatic gradient and diverse vegetation types, the threshold method has strong stability and adaptability, and has been widely used in alpine grassland ecosystems such as the Qinghai–Tibet Plateau [15,18]. Therefore, a dynamic threshold method was adopted in this study to extract SOS, EOS, and LOS, aiming to improve the flexibility and accuracy of phenological parameter extraction [46,47].
Threshold settings have a considerable impact on phenological results, and inappropriate threshold values may lead to systematic deviations in SOS and EOS detection. Previous studies have shown that in regions such as the Qinghai–Tibet Plateau, Northeast China, and North China, SOS is typically defined as 10–20% of the seasonal NDVI amplitude, while EOS is often defined as 40–50% [48,49,50]. Based on these findings and the spectral characteristics of grasslands in the study area, we adopted a stable threshold scheme of SOS = 0.2 and EOS = 0.5, ensuring scientific robustness, regional applicability, and temporal comparability of the extracted phenological parameters.

2.4. Estimation of Grassland GPP

The GPP data used in this study are derived from the PML V2 product, developed by the Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences. This dataset integrates an improved light use efficiency (LUE) model with multi-source remote sensing and meteorological drivers, utilizing a refined daily time-step approach. It has demonstrated high accuracy in global carbon flux assessments, particularly over alpine grassland ecosystems.
On the GEE platform, the PML V2 data were processed to calculate both annual total GPP and seasonal cumulative GPP over the southern slopes of the Qilian Mountains. The four seasons were defined as follows: spring (March–May), summer (June–August), autumn (September–November), and winter (December–February of the following year). GPP values were aggregated for each period by summing monthly GPP data, enabling the characterization of intra-annual variations in carbon assimilation capacity across the alpine grasslands.

2.5. Sen’s Slope Estimator

Sen’s slope estimator is a widely used non-parametric method for detecting trends in long-term time series, particularly in ecological and climate-related studies. Unlike traditional linear regression, it does not require the data to follow a normal distribution and is highly robust to outliers, making it especially suitable for time series with significant variability or missing values.
Originally proposed by P.K. Sen [51], this method calculates the slope between all possible pairs of data points in a time series (i.e., the difference in observed values divided by the time interval), and uses the median of these slopes as an estimate of the overall trend. Its simplicity and resistance to the influence of extreme values make it well-suited for analyzing environmental changes. In this study, Sen’s slope estimator was applied to quantify the trends in vegetation phenological metrics—namely the SOS, EOS, and LOS—as well as GPP over the period 2001–2020 across the grasslands of the southern slopes of the Qilian Mountains. The slope results at the pixel level reveal spatially explicit patterns of temporal change, offering a basis for subsequent analyses of phenology–carbon coupling and cluster-based regionalization.

2.6. Pixel-Wise Univariate Linear Regression Analysis

To quantitatively assess the response of carbon production potential to vegetation phenological changes in the grassland ecosystems of the southern slopes of the Qilian Mountains, this study employs pixel-wise univariate linear regression to model GPP as a function of phenological parameters (SOS, EOS, LOS) across spatial scales. This approach reveals the sensitivity and spatial heterogeneity of GPP responses to variations in phenological timing [52].
Specifically, phenological parameters (SOS, EOS, LOS) serve as independent variables, and GPP as the dependent variable, with regression models constructed at each pixel. Annual data from 2001 to 2020 are used to fit these models, yielding pixel-level slope coefficients (β) and significance levels (p-values). The slope β represents the magnitude and direction of GPP’s response to phenological changes. Specifically, β > 0 means that a delay in the phenological event (e.g., later SOS) is associated with an increase in GPP, whereas β < 0 means that an advance in the phenological event (e.g., earlier SOS) corresponds to an increase in GPP.
This regression analysis was implemented in the R programming environment (version 4.5.0), using pixel-level time series data to build the models and applying significance testing to identify regions with reliable responses. The method facilitates understanding the spatiotemporal patterns of carbon flux sensitivity to phenology in alpine grasslands and provides a quantitative foundation for subsequent clustering and mechanism analysis. For each pixel, three sensitivity coefficients were obtained: β_SOS, β_EOS and β_LOS, representing the responses of GPP to SOS, EOS and LOS, respectively.

2.7. K-Means Clustering

K-means clustering is an iterative, unsupervised learning algorithm widely used for partitioning large datasets into distinct groups. The clustering outcome strongly depends on the selection of the number of clusters (K), making the determination of an optimal K critical.
In this study, we combined the Silhouette coefficient and the Elbow method to select the best K value. The Silhouette coefficient measures the cohesion within clusters and the separation between clusters, with values closer to 1 indicating better-defined clusters [53]. The Elbow method involves plotting the Within-Cluster Sum of Squares (WSS) against different K values and identifying the “elbow” point where the rate of decrease sharply changes, signifying an optimal number of clusters [54].
To classify the coupled phenology–carbon production patterns of grasslands in the southern slopes of the Qilian Mountains, K-means clustering was performed on the standardized phenological variables—SOS, EOS, and LOS—exploring cluster numbers ranging from 2 to 10. The optimal K was determined by analyzing the WSS trend via the Elbow method and evaluating cluster stability and separation using the Silhouette coefficient, ensuring scientifically robust and ecologically meaningful classification results.

3. Results

3.1. Spatial and Temporal Distribution Characteristics of Grassland Phenology on the South Slope of Qilian Mountains

During the period 2001–2020, the grassland phenology on the southern slope of the Qilian Mountains exhibited clear seasonal patterns and notable spatial heterogeneity. The average SOS occurred on the 128th day of the year, primarily within April to May. The EOS averaged around the 264th day, mostly between September and October. The average LOS was approximately 135 days, ranging from 110 to 160 days (Figure 2), indicating that the regional growing season was relatively short and concentrated in the warm and moist period of the year.
In terms of temporal trends, the phenological parameters showed dynamic interannual variations. The SOS exhibited an advancing trend, with an average shift of 0.5 days per year, suggesting an earlier onset of vegetation greening. The LOS extended by approximately 0.6 days per year on average, indicating a gradual lengthening of the growing season. Meanwhile, the EOS was delayed by about 0.2 days per year, implying a slight postponement of vegetation senescence (Figure 3).
Further analysis based on the Sen’s slope estimator (Figure 2) revealed spatially heterogeneous trends across the study area. SOS trends in most regions ranged between −1 and 0 days per year, indicating a modest advancement of the growing season in most pixels. In the northwestern high-altitude zones, some localized areas showed stronger advancement trends ranging from −4 to −1 days per year, albeit with limited spatial extent. EOS trends were generally stable across space, with most areas showing slight advances between −1 and 0 days per year, and only a few pixels displaying delays. LOS showed broader spatial variability, with several areas experiencing annual extensions exceeding 1 day. However, areas with slight LOS shortening (−1 to 0 days per year) remained predominant, reflecting strong spatial heterogeneity in phenological responses.

3.2. Spatiotemporal Characteristics of Grassland GPP on the Southern Slope of the Qilian Mountains

From 2001 to 2020, the annual average GPP of grasslands on the southern slope of the Qilian Mountains exhibited a clear spatial gradient, decreasing from southeast to northwest, a pattern that closely aligns with the region’s elevation distribution (Figure 4). In lower-elevation areas, GPP values exceeded 1000 g C/m2/a in some locations, indicating strong photosynthetic potential.
GPP showed pronounced seasonal variability, with summer contributing the highest values, followed by spring and autumn, and the lowest values observed in winter. During summer, abundant solar radiation and favorable hydrothermal conditions enhanced photosynthetic activity, making it the peak season for carbon fixation. The spatial distribution of summer GPP mirrored that of the annual average, also showing a southeast-to-northwest decreasing trend. Spring and autumn GPP patterns were generally similar, with multi-year averages of 62.5 g C/m2/a and 96.3 g C/m2/a, respectively. In winter, GPP approached zero across most of the region due to average temperatures falling below 0 °C, which suppresses photosynthesis.
In terms of interannual variability (Figure 5), the annual GPP across the study area showed an overall increasing trend, with an average annual growth rate of 5.4 g C/m2/a. The peak value was observed in 2016, with an average GPP of 642.8 g C/m2/a. All four seasons exhibited increasing GPP trends, with the most substantial growth in summer (3.8 g C/m2/a), followed by autumn and spring.
Further analysis using the Sen’s slope estimator revealed spatial patterns in GPP trends. The spatial distribution of changes in summer and annual GPP was largely consistent, indicating strong spatial coupling. Annual GPP changes ranged from −49.26 to +41.33 g C/m2/a, while summer GPP varied from −34.13 to +32.5 g C/m2/a (Figure 6), highlighting summer’s dominant role in annual GPP dynamics. Spring, autumn, and winter GPP showed relatively moderate variations, ranging from −5.36 to +5.21, −10.18 to +6.55, and −7 to +4.61 g C/m2/a, respectively, with relatively stable fluctuations.
Overall, while the magnitude of GPP change differed by season, the spatial pattern revealed that the southeastern and central low-elevation areas experienced the most significant—and largely positive—changes. These trends may be closely related to regional warming, improved precipitation conditions, or vegetation recovery, reflecting an increasing carbon sink potential under ongoing climate change.

3.3. Sensitivity of GPP to Phenological Changes

To quantitatively assess the regulatory effects of vegetation phenology on carbon productivity on the southern slope of the Qilian Mountains, pixel-wise univariate linear regression analyses were conducted using annual GPP as the response variable and the start of season (SOS), end of season (EOS), and length of season (LOS) as explanatory variables. For each regression, the slope and corresponding p-value were extracted to characterize the magnitude, direction, and statistical significance of interannual GPP responses to phenological variability.
The spatial distribution of GPP sensitivity to SOS exhibited pronounced heterogeneity across the study area (Figure 7). In most regions, negative regression slopes dominated, indicating that delayed spring green-up was generally associated with reduced annual GPP, whereas earlier SOS tended to enhance carbon uptake. Statistically significant relationships were spatially sparse and fragmented, accounting for less than 5% of the total area, suggesting that although SOS–GPP coupling is widespread, its strength varies considerably across different environmental settings. Quantitatively, the regression slope of GPP with respect to SOS ranged from −14 to 13 g C m−2 yr−1 day−1, with most values concentrated between −2 and 2 g C m−2 yr−1 day−1 (Figure 8). Among these, the slope interval from −2 to −0.5 g C m−2 yr−1 day−1 accounted for the largest proportion (27.13%), and pixels with negative slopes represented approximately 59.23% of the study area.
For EOS, spatial patterns of GPP sensitivity were more heterogeneous and lacked a consistent directional trend (Figure 7). Both positive and negative responses were observed across the region, and statistically significant pixels were rare and spatially scattered, with a total proportion below 5%. This indicates that the influence of EOS on annual GPP is highly context-dependent. In terms of magnitude, EOS-related regression slopes ranged from −22 to 45 g C m−2 yr−1 day−1 and were primarily concentrated within the −2 to 2 interval (Figure 8), with roughly equal proportions of positive and negative slopes. Although a delayed EOS may promote higher GPP by extending the photosynthetically active period in some areas, this effect was offset or even reversed in certain low-elevation regions, possibly due to climatic constraints or physiological senescence.
Regarding LOS, the spatial distribution of GPP sensitivity revealed a predominance of positive responses across much of the study area (Figure 7), indicating that a longer growing season generally favors enhanced carbon productivity. However, negative responses were evident in some high-elevation regions, likely reflecting thermal and moisture limitations on alpine grasslands. The proportion of statistically significant LOS–GPP relationships (5.1%) was comparable to that observed for SOS. Quantitatively, regression slopes associated with LOS ranged from −11 to 17 g C m−2 yr−1 day−1, with approximately 50.01% of pixels falling within the narrow interval of −0.5 to 0.5 g C m−2 yr−1 day−1 (Figure 8). Overall, 56.22% of the study area exhibited positive slopes, suggesting that an extended growing season does not necessarily translate into increased carbon fixation capacity under all environmental conditions.

3.4. Coupling Patterns of Vegetation Phenology with GPP

To comprehensively identify the regulatory patterns of vegetation phenology on carbon productivity across the grasslands of the southern slope of the Qilian Mountains, we constructed a three-dimensional sensitivity feature space based on pixel-wise regression slopes of GPP against SOS, EOS, and LOS. K-means clustering was subsequently applied to classify the pixels according to their phenology–GPP coupling characteristics. Evaluation using the elbow method and silhouette coefficient indicated that K = 3 provided an optimal balance between clustering stability and ecological interpretability, with a pronounced reduction in within-cluster sum of squares and silhouette values exceeding 0.3 (Figure 9).
The resulting three phenology–GPP sensitivity clusters exhibit clearly differentiated functional characteristics. Cluster I, identified as a high carbon sequestration potential zone, is characterized by a negative sensitivity to delayed SOS (−0.6) and positive sensitivities to delayed EOS (1.9) and extended LOS (0.8). This pattern indicates that GPP in this zone benefits from a phenological configuration of earlier green-up, later senescence, and a prolonged growing season. Although clustering was based on GPP sensitivities rather than phenological trends themselves, ecosystems within Cluster I show a strong capacity to translate favorable phenological shifts into enhanced carbon uptake, reflecting high ecological adaptability.
Cluster II, defined as an ecologically vulnerable zone, exhibits uniformly negative sensitivity center values (SOS = −0.8, EOS = −3.4, LOS = −0.1), indicating that phenological shifts at any stage are detrimental to GPP. This response pattern suggests that, in these high-altitude ecosystems, phenological extensions do not translate into increased carbon uptake, but instead coincide with enhanced environmental stress. Warming-induced earlier snowmelt and green-up may advance the growing season into periods with limited soil moisture recharge, while delayed senescence can expose vegetation to increasing drought stress or early frost risk. In addition, dominant alpine grass species in these regions may exhibit limited physiological plasticity, constraining their ability to sustain photosynthetic activity over an extended growing season. As a result, phenological changes fail to enhance—and may even suppress—carbon assimilation in these regions.
Cluster III, categorized as a phenologically neutral buffer zone, displays sensitivity values close to zero (SOS = 0.1, EOS = −0.15, LOS = −0.1). Importantly, the extensive spatial dominance of this cluster, accounting for 55.2% of the study area (Figure 10), necessitates interpretation beyond a simplistic attribution to ecosystem stability. The near-zero sensitivities indicate a weak coupling between interannual GPP variability and the timing of season onset and termination; however, this widespread pattern more plausibly reflects a shift in dominant controls on carbon uptake. Specifically, GPP dynamics across large portions of the study area are likely regulated more strongly by mid-growing-season environmental conditions—such as the intensity and timing of peak-summer precipitation, soil moisture availability, and associated hydrothermal balance—than by variations in SOS or EOS. During the period of maximum photosynthetic demand, these mid-season factors can override phenological influences, effectively decoupling annual carbon assimilation from growing-season length. Consequently, the “neutral” response observed in Cluster III does not imply ecological insensitivity, but rather indicates the limitation of phenology-centric metrics in capturing the primary drivers of GPP variability in these regions.
Spatially, Cluster III is widely distributed across transitional mid-elevation zones, where hydrothermal conditions are relatively balanced and carbon uptake is less constrained by phenological boundaries. In contrast, Cluster I comprises 30.2% of pixels and is mainly located in southeastern and central low- to mid-elevation areas with favorable environmental conditions, while Cluster II accounts for 14.6% of the study area and is concentrated in high-altitude northwestern regions characterized by pronounced climatic stress. Together, this sensitivity-based clustering reveals strong functional heterogeneity in phenology–GPP coupling and highlights the existence of distinct regulatory regimes governing carbon dynamics across the southern slope of the Qilian Mountains.

4. Discussion

4.1. Spatiotemporal Patterns of Phenology and GPP in Alpine Grasslands

During the 2001–2020 study period, vegetation phenology on the southern slope of the Qilian Mountains exhibited a pronounced trend of earlier spring green-up, later autumn senescence, and an overall lengthening of the growing season, which are closely linked to global warming [55]. However, in alpine regions of the Qinghai–Tibet Plateau, increasing evidence from Earth observation studies indicates that the advancement of spring phenology is not solely controlled by rising air temperature, but is strongly mediated by changes in snow cover duration and snowmelt timing. Reduced snow persistence and earlier snowmelt can accelerate soil thawing and improve early-season soil moisture availability, thereby facilitating earlier vegetation green-up (SOS) [8,9].
Specifically, the SOS advanced by approximately 0.5 days per year, the EOS was delayed by about 0.2 days per year, and the LOS extended by an average of 0.6 days per year. Spatially, SOS showed a marked northwest–southeast advancement, while EOS was correspondingly delayed, resulting in a clear spatial gradient of growing-season lengthening [56,57]. Such spatial patterns are consistent with previous findings in the Qilian Mountains and across the Qinghai–Tibet Plateau, where elevation-dependent snow cover and snowmelt dynamics exert strong control on the timing of spring phenological onset [8,58].
Meanwhile, GPP of grassland on the southern slope has generally increased against the backdrop of a distinct warming and humidifying trend in this region [59]. The region’s complex terrain creates significant spatial heterogeneity in ecosystem processes. Existing studies indicate that the net effect of combined water and heat availability on regional grassland carbon productivity is positive [60,61], and our observed GPP increase is consistent with basin-wide studies showing the most pronounced increases occurring in summer [62,63]. The mean annual GPP for perennial grassland on the southern slope exceeded 530 g C m−2, slightly higher than the Qinghai–Tibet Plateau average [62], which likely reflects our focus on undisturbed perennial grassland, reducing uncertainty from mixed vegetation types.

4.2. Pixel-Scale Sensitivity of GPP to Phenological Shifts

This study developed pixel-scale linear regression models to evaluate the sensitivity of GPP responses to SOS, EOS, and LOS. The results reveal that these phenological metrics exert distinct effects on GPP, indicating spatial heterogeneity. Specifically, earlier SOS generally has positive effects on GPP, as earlier green-up enables an extended photosynthetic window and enhances early-season biomass accumulation. This aligns with findings that earlier spring phenology increases vegetation NPP through extended growing seasons [64,65]. Similarly, delayed EOS and increased LOS enhance GPP in some regions, as autumn extensions allow continued photosynthetic activity, contributing to carbon uptake, which is corroborated by studies on the Tibetan Plateau [66].
However, in high-altitude, climate-stressed areas, phenological advances or extensions may not lead to GPP increases and can even correlate negatively with productivity. This aligns with Zhang et al., who found that although green-up dates advanced across the plateau, NDVI did not increase in high-altitude zones—likely due to constraints such as low temperature, permafrost, and mismatches in water and heat that decouple potential growth from actual accumulation [67]. This paradox may stem from thermal–moisture mismatches or species-specific physiological constraints.

4.3. Ecological Implications of Phenology-GPP Functional Zonation

The three-dimensional sensitivity space combined with k-means clustering effectively delineates the functional heterogeneity of phenology–GPP coupling across alpine grasslands, revealing distinct ecological regimes in which carbon uptake responds differently to phenological variability.
Cluster I, characterized by high carbon sequestration potential, exhibits strong positive sensitivities of GPP to phenological changes. In these areas, extensions of the growing season—via earlier start of season (SOS) or delayed end of season (EOS)—translate directly into enhanced carbon assimilation. This pattern indicates that phenological shifts act as a primary regulator of interannual GPP variability, highlighting regions where climate-driven phenological advances can substantially strengthen ecosystem carbon sinks.
In contrast, Cluster II represents an ecologically vulnerable regime in which phenological extensions are associated with reductions in GPP. The observed “phenological extension–GPP reduction” paradox in these predominantly high-altitude regions reflects the dominance of physiological and environmental stress over phenological benefits. Earlier onset of the growing season driven by warming-induced snowmelt may lead to a thermal–moisture mismatch, whereby increased evaporative demand is not matched by sufficient soil moisture availability, resulting in progressive drought stress during the extended growing season. Moreover, species-specific physiological constraints may further limit carbon uptake. Dominant alpine grass species (e.g., Stipa purpurea) are often adapted to short growing seasons and conservative resource-use strategies, exhibiting limited plasticity in photosynthetic capacity, stomatal regulation, and carbon allocation. Consequently, an extended growing season may increase maintenance respiration costs without a corresponding increase in photosynthetic carbon gain. Finally, delayed senescence can prolong exposure to early autumn frost events, causing damage to photosynthetic tissues and further reducing seasonal carbon assimilation. Together, these mechanisms indicate that, in Cluster II, phenological lengthening amplifies environmental stress rather than enhancing productivity, resulting in net reductions in GPP.
The extensive spatial dominance of Cluster III (the so-called “neutral buffer” zone), accounting for 55.2% of the study area, requires careful interpretation beyond a simplistic attribution to ecosystem stability. While the near-zero sensitivity values indicate a weak statistical relationship between interannual GPP variability and the selected phenological metrics (SOS, EOS, and LOS), the large spatial coverage of this cluster strongly suggests an alternative explanation: that phenology may not be the primary control on carbon uptake dynamics across more than half of the alpine grasslands. Instead, GPP in these regions is likely more strongly regulated by environmental conditions during the mid-growing season, when photosynthetic activity reaches its maximum. Factors such as the magnitude and timing of peak-summer precipitation, soil moisture availability, and episodic heat stress can exert dominant control over GPP variability, effectively decoupling carbon assimilation from changes in the timing of seasonal onset and termination. Under such conditions, variations in SOS or EOS may have limited influence on annual carbon uptake, even though vegetation remains climatically responsive. This interpretation is consistent with the pronounced phenological sensitivity observed in Cluster I, underscoring a fundamental spatial dichotomy in the dominant regulatory mechanisms of carbon cycling across the landscape. Consequently, Cluster III should not be interpreted as an inherently “stable” ecosystem state, but rather as a region where a phenology-centric framework reveals its limitations, highlighting the necessity of incorporating mid-season hydro-meteorological drivers to fully explain observed GPP dynamics.
This functional zonation complements the multi-stage phenology–physiology framework proposed by Xia et al. [68], which emphasizes the combined control of GPP by phenological processes and vegetation functional traits. By explicitly identifying regions where GPP is governed primarily by non-phenological constraints, our results provide a clearer basis for interpreting the ecological meaning of the “neutral buffer” zone. From a management perspective, this distinction is critical: while carbon sink enhancement in Cluster I may largely depend on anticipating phenological shifts, strategies in Cluster III are more likely to benefit from managing water availability and mitigating climate extremes rather than focusing solely on changes in growing-season timing.

4.4. Limitations and Future Research Directions

This study provides a spatially explicit assessment of phenology–GPP coupling; however, several limitations should be acknowledged. First, the use of MODIS NDVI data at a 500 m spatial resolution, while suitable for regional-scale analysis, may smooth fine-scale ecological heterogeneity in the complex terrain of the Qilian Mountains. This spatial averaging effect could reduce sensitivity in topographic transition zones (ecotones) and potentially influence pixel-level estimates of phenology–GPP relationships.
Second, although the PML_V2 GPP product has been widely validated across diverse ecosystems, the lack of direct validation using in situ flux tower or phenological observations within the specific study area remains a limitation. This constraint is common across the Qinghai–Tibetan Plateau due to the sparse distribution of ground-based observation networks [44].
In addition, soil moisture and snow–ice processes are widely recognized as key regulators of alpine vegetation phenology and carbon uptake. While these factors were not explicitly included in the current sensitivity framework—which focuses on isolating the net effects of phenological timing (SOS, EOS, LOS)—their influence likely contributes to the observed spatial patterns, particularly in regions exhibiting weak or neutral phenology–GPP sensitivity. This interpretation is consistent with previous syntheses highlighting water availability as a primary constraint on warming-induced productivity responses on the Qinghai–Tibetan Plateau [44].
Future research should therefore prioritize a more integrated framework by (1) incorporating higher-resolution satellite data and ground-based observations (e.g., PhenoCams and eddy covariance flux towers) to improve spatial validation and (2) explicitly integrating satellite-derived soil moisture, snow cover duration, and land surface freeze–thaw indicators (e.g., from SMAP and SMOS missions) into phenology–carbon sensitivity analyses. Such efforts will be essential for disentangling the combined roles of phenological shifts and hydrothermal processes—particularly spring soil thawing and moisture availability—in regulating alpine grassland carbon dynamics.

5. Conclusions

Based on 2001–2020 satellite observations, this study provides a pixel-level, sensitivity-based assessment of the spatiotemporal coupling between vegetation phenology and gross primary productivity (GPP) in the alpine grasslands on the southern slope of the Qilian Mountains, a climatically sensitive but under-investigated transition zone on the northeastern Tibetan Plateau. By integrating MODIS-derived phenological metrics (SOS, EOS and LOS) with the advanced PML V2 GPP product at 500 m resolution, we revealed a consistent trend of earlier SOS, delayed EOS and prolonged LOS driven by earlier spring warming, later autumn cooling and increased heat accumulation. Meanwhile, GPP showed a steady upward trajectory, but its response to phenological shifts exhibited pronounced spatial heterogeneity along topographic and climatic gradients.
Earlier SOS generally enhanced carbon uptake, and in some regions, delayed EOS and extended LOS further promoted GPP. However, in high-elevation zones with strong climatic constraints, phenological advances or extensions did not necessarily translate into greater carbon accumulation, indicating a partial decoupling between longer growing seasons and effective carbon assimilation. These findings underscore the importance of considering both the timing and the effectiveness of phenological changes when evaluating carbon dynamics in alpine grasslands.
Building on pixel-wise sensitivity analysis of GPP to SOS, EOS and LOS, we constructed a three-dimensional phenology–GPP sensitivity space and applied k-means clustering to classify grassland ecosystems into three functional response zones—areas of high carbon sink potential, ecologically fragile regions and neutral buffers. Compared with previous studies that mainly described large-scale correlations between phenology and productivity, this sensitivity-based functional zoning framework explicitly captures how strongly and in which direction GPP responds to phenological shifts at the pixel scale. The resulting functional zones exhibit distinct phenology–productivity coupling patterns shaped by the combined influence of topography, climate variability and ecosystem resilience, and reveal the complex, spatially heterogeneous nature of phenology–carbon interactions in an ecological transition zone that has been underrepresented in earlier work.
Under the ongoing influence of global climate change, this study highlights the necessity of considering the entire growing season—encompassing SOS, EOS and LOS—in assessments of regional carbon dynamics and ecosystem services. The sensitivity-based and spatially explicit approach developed here offers practical guidance for identifying areas with high carbon sink potential and ecological vulnerability, providing scientific support for regionally differentiated climate adaptation and sustainable grassland management strategies on the southern slope of the Qilian Mountains and other alpine grasslands.
Future work should integrate in situ phenological observations to further refine remote sensing estimates, incorporate ecosystem models to simulate phenology–carbon coupling under diverse climate scenarios, and explore non-climatic drivers such as soil moisture and freeze–thaw processes. These steps will strengthen our mechanistic understanding and improve projections of alpine grassland carbon sequestration in a rapidly changing environment.

Author Contributions

Conceptualization, F.W., Y.Z. and M.Z.; Methodology, F.W., Y.Z., Y.W. and M.Z.; Validation, F.W. and M.Z.; Formal analysis, F.W.; Writing—original draft preparation, Y.Z. Writing—review and editing, F.W. and M.Z.; Supervision, M.Z.; Project administration & Funding acquisition, G.C. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the Natural Science Foundation of the Qinghai Province (grant no. 2023-ZJ-907M).

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The original contributions presented in this study are included in the article.

Conflicts of Interest

The authors declare no conflicts of interest.

Abbreviations

The following abbreviations are used in this manuscript:
SOSStart of the growing season
LOSLength of growing season
EOSEnd of the growing season
GPPGross primary productivity

References

  1. Kotlarski, S.; Gobiet, A.; Morin, S.; Olefs, M.; Rajczak, J.; Samacoïts, E. 21st Century alpine climate change. Clim. Dyn. 2023, 60, 65–86. [Google Scholar] [CrossRef]
  2. Li, L.; Zhang, Y.; Wu, J.; Li, S.; Zhang, B.; Zu, J.; Zhang, H.; Ding, M.; Paudel, B. Increasing sensitivity of alpine grasslands to climate variability along an elevational gradient on the Qinghai-Tibet Plateau. Sci. Total Environ. 2019, 678, 21–29. [Google Scholar] [CrossRef]
  3. Jin, J.; Wang, Y.; Zhang, Z.; Magliulo, V.; Jiang, H.; Cheng, M. Phenology Plays an Important Role in the Regulation of Terrestrial Ecosystem Water-Use Efficiency in the Northern Hemisphere. Remote Sens. 2017, 9, 664. [Google Scholar] [CrossRef]
  4. Yang, F.; Liu, C.; Chen, Q.; Lai, J.; Liu, T. Earlier Spring-Summer Phenology and Higher Photosynthetic Peak Altered the Seasonal Patterns of Vegetation Productivity in Alpine Ecosystems. Remote Sens. 2024, 16, 1580. [Google Scholar] [CrossRef]
  5. Kang, X.; Hao, Y.; Cui, X.; Chen, H.; Huang, S.; Du, Y.; Li, W.; Kardol, P.; Xiao, X.; Cui, L. Variability and Changes in Climate, Phenology, and Gross Primary Production of an Alpine Wetland Ecosystem. Remote Sens. 2016, 8, 391. [Google Scholar] [CrossRef]
  6. Yuan, M.; Wen, Z.; He, L.; Li, X.; Zhao, L. Response of temperate grassland phenology to climate change and its contribution to gross primary productivity in China. Acta Ecol. Sin. 2024, 44, 354–376. [Google Scholar]
  7. Zhang, Y.; Wang, H.; Gao, C.; Lin, Z.; Zhou, W. Vegetation greenness and photosynthetic spring phenology in Northeast China and their responses to water and thermal factors. Prog. Geogr. 2024, 43, 2507–2519. [Google Scholar]
  8. Qi, Y.; Wang, H.; Ma, X.; Zhang, J.; Yang, R. Relationship between vegetation phenology and snow cover changes during 2001–2018 in the Qilian Mountains. Ecol. Indic. 2021, 133, 108351. [Google Scholar] [CrossRef]
  9. Ma, X.; Zhang, Y.; Wang, J. Snow depth dominates spring phenology of alpine vegetation on the Tibetan Plateau. Agric. For. Meteorol. 2025, 360, 109208. [Google Scholar]
  10. Liao, Z.; Zhou, B.; Zhu, J.; Jia, H.; Fei, X. A critical review of methods, principles and progress for estimating the gross primary productivity of terrestrial ecosystems. Front. Environ. Sci. 2023, 11, 1093095. [Google Scholar] [CrossRef]
  11. Gómez-Giráldez, P.J.; Cristóbal, J.; Nieto, H.; García-Díaz, D.; Díaz-Delgado, R. Validation of Gross Primary Production Estimated by Remote Sensing for the Ecosystems of Doñana National Park through Improvements in Light Use Efficiency Estimation. Remote Sens. 2024, 16, 2170. [Google Scholar] [CrossRef]
  12. Yin, H.; Ma, X.; Liao, X.; Ye, H.; Yu, W.; Li, Y.; Wei, J.; Yuan, J.; Liu, Q. Linking Vegetation Phenology to Net Ecosystem Productivity: Climate Change Impacts in the Northern Hemisphere Using Satellite Data. Remote Sens. 2024, 16, 4101. [Google Scholar] [CrossRef]
  13. Zhang, J.; Xiao, J.; Tong, X.; Zhang, J.; Meng, P.; Li, J.; Liu, P.; Yu, P. NIRv and SIF better estimate phenology than NDVI and EVI: Effects of spring and autumn phenology on ecosystem production of planted forests. Agric. For. Meteorol. 2022, 315, 108819. [Google Scholar] [CrossRef]
  14. Sun, Y.; Guan, Q.; Du, Q.; Wang, Q.; Yang, X.; Zhang, E. Effects of spring phenology on grassland growth in Qilian Mountains across multiple spatiotemporal scales. JGR Biogeosci. 2023, 128, e2023JG007557. [Google Scholar] [CrossRef]
  15. Li, C.; Zou, Y.; He, J.; Zhang, W.; Gao, L.; Zhuang, D. Response of Vegetation Phenology to the Interaction of Temperature and Precipitation Changes in Qilian Mountains. Remote Sens. 2022, 14, 1248. [Google Scholar] [CrossRef]
  16. Zhao, J.; Wang, Y.; Zhang, Z.; Zhang, H.; Guo, X.; Yu, S.; Du, W.; Huang, F. The Variations of Land Surface Phenology in Northeast China and Its Responses to Climate Change from 1982 to 2013. Remote Sens. 2016, 8, 400. [Google Scholar] [CrossRef]
  17. Xu, L.; Niu, B.; Zhang, X.; He, Y. Dynamic Threshold of Carbon Phenology in Two Cold Temperate Grasslands in China. Remote Sens. 2021, 13, 574. [Google Scholar] [CrossRef]
  18. Zhang, Y.; Cao, G.; Zhao, M.; Zhang, Q.; Huang, L. Integrated Effects of Climate, Topography, and Greenhouse Gas on Grassland Phenology in the Southern Slope of the Qilian Mountains. Atmosphere 2025, 16, 653. [Google Scholar] [CrossRef]
  19. Qiao, C.; Shen, S.; Cheng, C.; Wu, J.; Jia, D.; Song, C. Vegetation Phenology in the Qilian Mountains and Its Response to Temperature from 1982 to 2014. Remote Sens. 2021, 13, 286. [Google Scholar] [CrossRef]
  20. Zhou, G.; Ren, H.; Zhang, L.; Lv, X.; Zhou, M. Annual vegetation maps in the Qinghai–Tibet Plateau (QTP) from 2000 to 2022 based on MODIS series satellite imagery. Earth Syst. Sci. Data 2025, 17, 773–797. [Google Scholar] [CrossRef]
  21. Li, X.; Zhao, C.; Kang, M.; Ma, M. Responses of net primary productivity to phenological dynamics based on a data fusion algorithm in the northern Qinghai–Tibet Plateau. Ecol. Indic. 2022, 142, 109239. [Google Scholar] [CrossRef]
  22. Liu, H.; Liu, S.; Wang, F. Management practices should be strengthened in high potential vegetation productivity areas based on vegetation phenology assessment on the Qinghai–Tibet Plateau. Ecol. Indic. 2022, 140, 108991. [Google Scholar] [CrossRef]
  23. Liu, Y.; Zhang, X.; Sun, M.; Du, X.; Zhu, Q. Widespread increasing negative vegetation sensitivity to phenology reshapes GPP dynamics in alpine grassland. J. Environ. Manag. 2025, 392, 126680. [Google Scholar] [CrossRef] [PubMed]
  24. Liu, S.; Zhang, F.; Du, Y.; Guo, X.; Lin, L.; Li, Y.; Li, Q.; Cao, G. Ecosystem Carbon Storage in Alpine Grassland on the Qinghai Plateau. PLoS ONE 2016, 11, e0160420. [Google Scholar] [CrossRef]
  25. Wang, Y.; Xiao, J.; Ma, Y.; Ding, J.; Chen, X.; Ding, Z.; Luo, Y. Persistent and enhanced carbon sequestration capacity of alpine grasslands on Earth’s Third Pole. Sci. Adv. 2023, 9, eade6875. [Google Scholar] [CrossRef]
  26. Huo, Q.; Ma, K.; Yu, X. Interannual variations in grassland carbon fluxes and attribution of influencing factors in Qilian Mountains, China. Sci. Total Environ. 2024, 957, 177786. [Google Scholar]
  27. Abalori, T.A.; Cao, W.; Atogi-Akwoa Weobong, C.; Sam, F.E.; Li, W.; Osei, R.; Wang, S. Effects of vegetation patchiness on ecosystem carbon and nitrogen storage in the alpine grassland of the Qilian Mountains. Front. Environ. Sci. 2022, 10, 879717. [Google Scholar] [CrossRef]
  28. Bai, Y.; Li, S. Growth peak of vegetation and its response to drought on the Mongolian Plateau. Ecol. Indic. 2022, 141, 109150. [Google Scholar] [CrossRef]
  29. He, P.; Ma, X.; Sun, Z. Interannual variability in summer climate change controls GPP long-term changes. Environ. Res. 2022, 212, 113409. [Google Scholar] [CrossRef]
  30. Zhang, Z.; Zhang, Z.; Hautier, Y.; Qing, H.; Yang, J.; Bao, T.; Hajek, O.L.; Knapp, A.K. Effects of intra-annual precipitation patterns on grassland productivity moderated by the dominant species phenology. Front. Plant Sci. 2023, 14, 1142786. [Google Scholar] [CrossRef]
  31. Balzarolo, M.; Peñuelas, J.; Veroustraete, F. Influence of landscape heterogeneity and spatial resolution in Multi-Temporal in situ and MODIS NDVI data proxies for seasonal GPP dynamics. Remote Sens. 2019, 11, 1656. [Google Scholar] [CrossRef]
  32. Zhang, Y.; Xiao, X.; Wu, X.; Zhou, S.; Zhang, G.; Qin, Y.; Dong, J. A global moderate resolution dataset of gross primary production of vegetation for 2000–2016. Sci. Data 2017, 4, 170165. [Google Scholar] [CrossRef] [PubMed]
  33. He, H.; Liu, M.; Xiao, X.; Ren, X.; Zhang, L.; Sun, X.; Yang, Y.; Li, Y.; Zhao, L.; Shi, P.; et al. Large-scale estimation and uncertainty analysis of gross primary production in Tibetan alpine grasslands. JGR Biogeosci. 2014, 119, 466–486. [Google Scholar] [CrossRef]
  34. He, S.; Zhang, Y.; Ma, N.; Tian, J.; Kong, D.; Liu, C. A daily and 500 m coupled evapotranspiration and gross primary production product across China during 2000–2020. Earth Syst. Sci. Data 2022, 14, 5463–5488. [Google Scholar] [CrossRef]
  35. Liu, Y.; Chen, C.; Chen, Q.; Zhang, J.; Cui, Z.; Zeng, Z. Widespread sensitivity of grassland water use efficiency to deep soil moisture on the Tibetan Plateau. Water Resour. Res. 2025, 61, e2024WR038645. [Google Scholar] [CrossRef]
  36. Zhang, Y.; Kong, D.; Gan, R.; Chiew, F.H.; McVicar, T.R.; Zhang, Q.; Yang, Y. Coupled estimation of 500 m and 8-day resolution global evapotranspiration and gross primary production in 2002–2017. Remote Sens. Environ. 2019, 222, 165–182. [Google Scholar] [CrossRef]
  37. Gan, R.; Zhang, Y.; Shi, H.; Yang, Y.; Eamus, D.; Cheng, L.; Chiew, F.H.; Yu, Q. Use of satellite leaf area index estimating evapotranspiration and gross assimilation for Australian ecosystems. Ecohydrology 2018, 11, e1974. [Google Scholar] [CrossRef]
  38. Ma, D.; Wu, X.; Ma, X.; Wang, J.; Lin, X.; Mu, C. Spatial, Phenological, and Inter-Annual Variations of Gross Primary Productivity in the Arctic from 2001 to 2019. Remote Sens. 2021, 13, 2875. [Google Scholar] [CrossRef]
  39. Duan, H.; Qi, Y.; Kang, W.; Zhang, J.; Wang, H.; Jiang, X. Seasonal Variation of Vegetation and Its Spatiotemporal Response to Climatic Factors in the Qilian Mountains, China. Sustainability 2022, 14, 4926. [Google Scholar] [CrossRef]
  40. Tang, Z.; Yu, Q.; Liu, H.; Jiang, S.; He, F.; Zhang, Y.; Wang, F.; Zhang, Y.; Zhao, H.; Zhao, P. Characteristics of alpine vegetation community and its relationship to topographic climate factors in the eastern Qilian mountain. Acta Ecol. Sin. 2020, 40, 223–232. [Google Scholar] [CrossRef]
  41. Wu, X.; Jiao, L.; Du, D.; Xue, R.; Ding, X.; Wei, M.; Zhang, P. Spatial–Temporal Pattern and Influencing Factors of Vegetation Phenology and Net Primary Productivity in the Qilian Mountains of Northwest China. Sustainability 2022, 14, 14337. [Google Scholar] [CrossRef]
  42. Tong, S.; Cao, G.; Yan, X.; Diao, E.; Zhang, Z. Spatial-Temporal evolution of vegetation cover and its driving factors on the South Slope of the Qilian Mountains, China from 2000 to 2020. Mt. Res. 2022, 40, 491–503. [Google Scholar]
  43. Zhang, Q.; Cao, G.; Zhang, L.; Zhao, M. Spatiotemporal changes in vegetation greenness on the southern slopes of the Qilian Mountains and their responses to climate change and human activities. Arid Zone Res. 2024, 41, 2143–2153. [Google Scholar]
  44. Wang, Y.; Lv, W.; Xue, K. Grassland changes and adaptive management on the Qinghai–Tibetan Plateau. Nat. Rev. Earth Environ. 2022, 3, 668–683. [Google Scholar] [CrossRef]
  45. Wang, C.; Li, J.; Liu, Q.; Zhong, B.; Wu, S.; Xia, C. Analysis of differences in phenology extracted from the enhanced vegetation index and the leaf area index. Sensors 2017, 17, 1982. [Google Scholar] [CrossRef]
  46. Wang, C.; Wang, J.; Wang, X.; Luo, D.; Wei, Y.; Cui, X.; Wu, N.; Bagaria, P. Phenological changes in alpine grasslands and their influencing factors in seasonally frozen ground regions across the three parallel rivers region, Qinghai-Tibet Plateau. Front. Earth Sci. 2022, 9, 797928. [Google Scholar] [CrossRef]
  47. Chen, T.; Chen, Z.; Xie, G. Spatiotemporal analysis of phenological metrics on the Qinghai-Tibet Plateau based on multiple vegetation indices. Front. Environ. Sci. 2024, 12, 1489267. [Google Scholar] [CrossRef]
  48. Jin, Z.; Zhuang, Q.; Dukes, J.S.; He, J.; Sokolov, A.P.; Chen, M.; Zhang, T.; Luo, T. Temporal variability in the thermal requirements for vegetation phenology on the Tibetan plateau and its implications for carbon dynamics. Clim. Change 2016, 138, 617–663. [Google Scholar] [CrossRef]
  49. Zhou, Y.; Jia, G. Precipitation as a control of vegetation phenology for temperate steppes in China. Atmos. Ocean. Sci. Lett. 2016, 9, 162–168. [Google Scholar] [CrossRef]
  50. Ji, Y.; Zhan, W.; Du, H.; Wang, S.; Li, L.; Xiao, J.; Liu, Z.; Huang, F.; Jin, J. Urban-rural gradient in vegetation phenology changes of over 1500 cities across China jointly regulated by urbanization and climate change. ISPRS J. Photogramm. 2023, 205, 367–384. [Google Scholar] [CrossRef]
  51. Sen, P.K. Estimates of the Regression Coefficient Based on Kendall’s Tau. J. Am. Stat. Assoc. 1986, 63, 1379–1389. [Google Scholar] [CrossRef]
  52. Zhao, S.; Wang, Y.; Qiao, X.; Zhao, T. Spatiotemporal variation and driving factors for FVC in Huaihe River Basin from 1987 to 2021. Trans. Chin. Soc. Agric. Mach. 2023, 54, 180–190. [Google Scholar]
  53. Jiao, L.; Wang, Y.; Ma, M.; Xu, Z. Multi-dimensional characteristics and pattern recognition of urban form: A Case Study of Natural Cities in China, the United States and Europe. Geomat. Inf. Sci. Wuhan Univ. 2024, 49, 1005–1017. [Google Scholar]
  54. Jin, H.; Chen, X.; Wu, P.; Song, C.; Xia, W. Evaluation of spatial-temporal distribution of precipitation in mainland China by statistic and clustering methods. Atmos. Res. 2021, 262, 105772. [Google Scholar] [CrossRef]
  55. Piao, S.; Liu, Q.; Chen, A.; Janssens, I.A.; Fu, Y.; Dai, J.; Liu, L.; Lian, X.; Shen, M.; Zhu, X. Plant phenology and global climate change: Current progresses and challenges. Glob. Chang. Biol. 2019, 25, 1922–1940. [Google Scholar] [CrossRef]
  56. Qiao, C.; Jia, D.; Cheng, C. Vegetation phenology change and response to temperature in the Qilian Mountains from 1982 to 2014. J. Beijing Norm. Univ. (Nat. Sci.) 2022, 58, 168–177. [Google Scholar]
  57. Jia, W.; Zhao, Z.; Zu, J.; Chen, J.; Wang, J.; Ding, D. Phenological variation in different vegetation types and their response to climate change in the Qilian Mountains, China, 1982–2014. Acta Ecol. Sin. 2016, 36, 7826–7840. [Google Scholar]
  58. Wang, S.; Zhang, B.; Yang, Q. Responses of vegetation phenology to climate change on the Tibetan Plateau: The role of snow cover. Remote Sens. Environ. 2021, 256, 112307. [Google Scholar]
  59. Yang, F.; Zhang, W.; Zhang, F.; Wang, C. Climate characteristics and variation in the Qilian Mountains from 1961 to 2022. Arid Zone Res. 2024, 41, 1627–1638. [Google Scholar]
  60. Wang, Y.; Liu, Y.; Zhou, L.; Zhou, G. Spatiotemporal patterns of phenological metrics and their relationships with environmental drivers in grasslands. Sci. Total Environ. 2024, 938, 173489. [Google Scholar] [CrossRef]
  61. Wei, D.; Zhang, Y.; Li, Y.; Zhang, Y.; Wang, B. Hydrothermal Conditions in Deep Soil Layer Regulate the Interannual Change in Gross Primary Productivity in the Qilian Mountains Area, China. Forests 2023, 14, 2422. [Google Scholar] [CrossRef]
  62. Lv, J.; Zhao, W. Variations of vegetation gross primary productivity and its driving factors in Tibetan Plateau. Acta Ecol. Sin. 2025, 45, 6934–6947. [Google Scholar]
  63. Xie, Z.; Li, Z. Temporal and spatial variations of gross primary productivity (GPP) in vegetation ecosystems and its dominant climatic factors identification in Southwest China. Guihaia 2025, 45, 1380–1391. [Google Scholar]
  64. Wang, X.; Xing, W. Analysis of the spatial and temporal characteristics of GPP of grassland vegetation in China and its influencing factors based on MODIS. Rural Econ. Sci. Technol. 2020, 31, 15–16. [Google Scholar]
  65. Piao, S.; Friedlingstein, P.; Ciais, P.; Viovy, N.; Demarty, J. Growing season extension and its impact on terrestrial carbon cycle in the Northern Hemisphere over the past 2 decades. Glob. Biogeochem. Cycles 2007, 21, GB3018. [Google Scholar] [CrossRef]
  66. Zheng, Z.; Zhu, W.; Zhang, Y. Direct and Lagged Effects of Spring Phenology on Net Primary Productivity in the Alpine Grasslands on the Tibetan Plateau. Remote Sens. 2020, 12, 1223. [Google Scholar] [CrossRef]
  67. Zhang, G.; Zhang, Y.; Dong, J.; Xiao, X. Green-up dates in the Tibetan Plateau have continuously advanced from 1982 to 2011, but NDVI has not increased at high elevations. Proc. Natl. Acad. Sci. USA 2013, 110, 4309–4314. [Google Scholar] [CrossRef]
  68. Xia, J.; Niu, S.; Ciais, P.; Luo, Y. Joint control of terrestrial gross primary productivity by plant phenology and physiology. Ecology 2015, 112, 2788–2793. [Google Scholar] [CrossRef]
Atmosphere 17 00169 g001
Figure 1. Geographic Location of the Study Area. ((a). Location of Qinghai Province within China; (b). Location of the southern slope of the Qilian Mountains within Qinghai Province; (c). Elevation of the southern slope of the Qilian Mountains).
Figure 1. Geographic Location of the Study Area. ((a). Location of Qinghai Province within China; (b). Location of the southern slope of the Qilian Mountains within Qinghai Province; (c). Elevation of the southern slope of the Qilian Mountains).
Atmosphere 17 00169 g001
Atmosphere 17 00169 g002
Figure 2. Multi-year Mean Values and Trend Analysis of Phenological Indicators. ((a). Start of season (SOS); (b). Trend of SOS; (c). End of season (EOS); (d). Trend of EOS; (e). Length of season (LOS); (f). Trend of LOS).
Figure 2. Multi-year Mean Values and Trend Analysis of Phenological Indicators. ((a). Start of season (SOS); (b). Trend of SOS; (c). End of season (EOS); (d). Trend of EOS; (e). Length of season (LOS); (f). Trend of LOS).
Atmosphere 17 00169 g002
Atmosphere 17 00169 g003
Figure 3. Interannual Variation Curves of Phenological Indicators. ((a). Start of season (SOS); (b). Length of season (LOS); (c). End of season (EOS)).
Figure 3. Interannual Variation Curves of Phenological Indicators. ((a). Start of season (SOS); (b). Length of season (LOS); (c). End of season (EOS)).
Atmosphere 17 00169 g003
Atmosphere 17 00169 g004
Figure 4. Multi-year Mean Spatial Distribution of Annual and Seasonal GPP in Qilian Mountains’ Southern Slope Grasslands. ((a). Spring; (b). Summer; (c). Autumn; (d). Winter; (e). Annual).
Figure 4. Multi-year Mean Spatial Distribution of Annual and Seasonal GPP in Qilian Mountains’ Southern Slope Grasslands. ((a). Spring; (b). Summer; (c). Autumn; (d). Winter; (e). Annual).
Atmosphere 17 00169 g004
Atmosphere 17 00169 g005
Figure 5. Interannual Variation Curves of Annual and Seasonal GPP in Qilian Mountains’ Southern Slope Grasslands. ((a). Seasonal GPP in spring, summer, autumn, and winter; (b). Annual GPP; (c). Spring GPP; (d). Summer GPP; (e). Autumn GPP; (f). Winter GPP).
Figure 5. Interannual Variation Curves of Annual and Seasonal GPP in Qilian Mountains’ Southern Slope Grasslands. ((a). Seasonal GPP in spring, summer, autumn, and winter; (b). Annual GPP; (c). Spring GPP; (d). Summer GPP; (e). Autumn GPP; (f). Winter GPP).
Atmosphere 17 00169 g005
Atmosphere 17 00169 g006
Figure 6. Multi-year Trends of Annual and Seasonal GPP Based on Sen’s Slope Estimator. ((a). Spring; (b). Summer; (c). Autumn; (d). Winter; (e). Annual).
Figure 6. Multi-year Trends of Annual and Seasonal GPP Based on Sen’s Slope Estimator. ((a). Spring; (b). Summer; (c). Autumn; (d). Winter; (e). Annual).
Atmosphere 17 00169 g006
Atmosphere 17 00169 g007
Figure 7. The distribution of response intensity and significance of GPP to phenological indicators. ((a,c,e). GPP sensitivity to SOS, EOS, and LOS, respectively; (b,d,f). Corresponding statistical significance).
Figure 7. The distribution of response intensity and significance of GPP to phenological indicators. ((a,c,e). GPP sensitivity to SOS, EOS, and LOS, respectively; (b,d,f). Corresponding statistical significance).
Atmosphere 17 00169 g007
Atmosphere 17 00169 g008
Figure 8. Percentage Bar Chart of GPP Response to Phenological Indicators (Categories 1–6 Correspond to Legend Colors in Figure 8). where values 1–3 are negative and values 4–6 are positive.
Figure 8. Percentage Bar Chart of GPP Response to Phenological Indicators (Categories 1–6 Correspond to Legend Colors in Figure 8). where values 1–3 are negative and values 4–6 are positive.
Atmosphere 17 00169 g008
Atmosphere 17 00169 g009
Figure 9. Combined Evaluation of Clustering Performance Using the Elbow Method and Silhouette Coefficient. ((a). Average silhouette coefficient for different numbers of clusters (K); (b). Within-cluster sum of squares (WSS) for different numbers of clusters (K)).
Figure 9. Combined Evaluation of Clustering Performance Using the Elbow Method and Silhouette Coefficient. ((a). Average silhouette coefficient for different numbers of clusters (K); (b). Within-cluster sum of squares (WSS) for different numbers of clusters (K)).
Atmosphere 17 00169 g009
Atmosphere 17 00169 g010
Figure 10. Spatial Distribution and Proportion of Grassland Clusters on the Southern Slope of the Qilian Mountains.
Figure 10. Spatial Distribution and Proportion of Grassland Clusters on the Southern Slope of the Qilian Mountains.
Atmosphere 17 00169 g010
Table 1. Data sources.
Table 1. Data sources.
DatasetTemporal CoverageSpatial ResolutionPurposeAccess Method
MCD12Q1 Land Cover Product2001–2020500 mIdentification of stable grassland areashttps://earthengine.google.com/ (accessed on 15 March 2025)
MOD13Q1 NDVI2001–2020500 mConstruction of NDVI time series and phenology extraction (SOS, EOS, LOS)https://earthengine.google.com/ (accessed on 20 March 2025)
PML V2 GPP Product2001–2020500 mEstimation of GPP and coupling analysis with phenologyhttps://earthengine.google.com/ (accessed on 1 April 2025)
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Wang, F.; Zhang, Y.; Cao, G.; Zhao, M.; Wang, Y. Coupling Mechanisms Between Vegetation Phenology and Gross Primary Productivity in Alpine Grasslands on the Southern Slope of the Qilian Mountains. Atmosphere 2026, 17, 169. https://doi.org/10.3390/atmos17020169

AMA Style

Wang F, Zhang Y, Cao G, Zhao M, Wang Y. Coupling Mechanisms Between Vegetation Phenology and Gross Primary Productivity in Alpine Grasslands on the Southern Slope of the Qilian Mountains. Atmosphere. 2026; 17(2):169. https://doi.org/10.3390/atmos17020169

Chicago/Turabian Style

Wang, Fangyu, Yi Zhang, Guangchao Cao, Meiliang Zhao, and Yinggui Wang. 2026. "Coupling Mechanisms Between Vegetation Phenology and Gross Primary Productivity in Alpine Grasslands on the Southern Slope of the Qilian Mountains" Atmosphere 17, no. 2: 169. https://doi.org/10.3390/atmos17020169

APA Style

Wang, F., Zhang, Y., Cao, G., Zhao, M., & Wang, Y. (2026). Coupling Mechanisms Between Vegetation Phenology and Gross Primary Productivity in Alpine Grasslands on the Southern Slope of the Qilian Mountains. Atmosphere, 17(2), 169. https://doi.org/10.3390/atmos17020169

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop