Abstract. Active layer moisture (ALM) plays a fundamental role in regulating the hydrothermal dynamics, freeze–thaw processes, carbon cycling, and ecosystem and load-bearing functions of permafrost. However, spatially continuous, high-resolution datasets that characterize soil moisture across the entire active layer remain largely unavailable for the Qinghai–Tibet Plateau (QTP), which hosts the largest low- to mid-latitude permafrost region globally. Here, we present the first 90 m resolution, spatially complete dataset of ALM – the depth-averaged volumetric water content of the entire active layer – for the entire permafrost region of the QTP. The dataset was produced by integrating 342 in situ samples collected during peak-thaw seasons from 2009 to 2024 with multi-source remote sensing environmental predictors, topographic factors, and soil physicochemical properties. Four ensemble machine learning models, include Random Forest, Extra Trees, XGBoost, and CatBoost, were trained, and a bias-aware multi-model fusion strategy was applied to generate the final ALM product. SHAP-based recursive feature elimination was used to identify the dominant environmental controls and optimize feature selection. Random five-fold cross-validation showed high apparent accuracy (R² =0.62–0.63, RMSE ≈ 0.08 m³ m⁻³), and a group-based spatial cross-validation provided a conservative transferability estimate (pooled R² = 0.30–0.38). Within the permafrost region, ALM ranges from 0.02 to 0.70 m³ m⁻³, with a mean of 0.21 m³ m⁻³ and a standard deviation of 0.09 m³ m⁻³, exhibiting a distinct southeast-to-northwest decreasing gradient. The dataset is freely available with DOI: https://doi.org/10.12072/ncdc.permafrost.db7312.2026. The dataset is accompanied by an area-of-applicability (AOA) mask delineating the region – approximately 60 % of the permafrost region – where the cross-validated performance can be expected to hold, together with per-pixel uncertainty estimates derived from the DI–RMSE relationship and the divergence among the four models.

The Tibetan Plateau (TP), known as the ‘Asian Water Tower’, poses significant challenges for hydrological modeling due to its complex cryospheric processes and parametric uncertainties. To address these challenges, we developed an integrated evaluation framework that combines spatiotemporal performance metrics with three global sensitivity analysis methods, based on the Variable Infiltration Capacity (VIC) model. Four alpine river basins were used as case studies to assess the impact of 33 parameters on nine hydro-energy variables across daily, seasonal, and spatial scales. Key drivers of spatial heterogeneity in parameter sensitivities were identified. The results indicate that snow albedo, leaf area index, and the soil drainage parameter broadly influence multiple processes. Runoff and baseflow sensitivities vary spatiotemporally. Random Forest-based SHapley Additive Explanations analysis reveals an east-west gradient in parameter sensitivity, driven by temperature, precipitation, and radiation. Two-step parameter optimization improves the average daily simulation efficiency and spatial consistency for land surface temperature and snow cover fraction by 33% and 30%, respectively, without compromising runoff accuracy. Transferring parameter sensitivities to similar basins confirms the framework's robustness and generalizability. This study underscores the importance of non-runoff parameters, enhances simulation performance, and provides insights into seasonal hydrological variability for more robust model applications across the TP.

Snow depth is a crucial parameter for describing the spatiotemporal variations of snow cover, and passive microwave snow depth products (10-25 km) are widely used for monitoring snow depth changes. However, as one of the three major snow-covered regions in China, the Qinghai-Tibet Plateau has complex terrain and rapid changes in snow cover with strong spatial heterogeneity, making it difficult for coarse-resolution snow depth products to accurately describe its spatiotemporal characteristics. This study proposes a high spatial resolution (500 m) snow depth estimation method based on the Advanced Microwave Scanning Radiometer 2 (AMSR-2) brightness temperature data and Automated Machine Learning. Firstly, using Pearson correlation coefficients, 19 key factors influencing snow depth, including AMSR-2 brightness temperature, slope, and surface roughness, were selected as input data (independent variables) for Automated Machine Learning. Meanwhile, passive microwave downscaled snow depth data and ground-based snow depth measurements were introduced as dependent variables for Automated Machine Learning. The Automated Machine Learning model was then trained separately for four different types of snow cover surfaces (forest, grassland, water, and bare land). Finally, through ten-fold cross-validation, the optimal machine learning model for snow depth estimation under each type of underlying surface coverage was selected, thus generating sequential snow depth datasets for the ten-year snow cover period of the Qinghai-Tibet Plateau from 2012 to 2021. Results show that (1) the estimated snow depth values well with ground-based observations, yielding a coefficient of determination (R-2) of 0.71 and a root mean square error (RMSE) of 3.64 cm, indicating high estimation accuracy. (2) Snow depth estimation demonstrates the highest accuracy in unused land (CatBoost, R-2 = 0.82), followed by grassland (CatBoost, R-2 = 0.77, RMSE = 3.11 cm), water (ET, R-2 = 0.75, RMSE = 2.20 cm), and forest (XGBoost, R-2 = 0.71, RMSE = 3.30 cm). (3) A comparison with snow cover extent derived from Landsat-8 optical imagery reveals that the estimated snow depth spatial distribution is consistent with snow cover extent, providing reliable data for monitoring snow cover changes in mountainous regions.

Significant uncertainties persist in current snow depth (SD) datasets due to variations in sensor characteristics and retrieval algorithms. This study systematically evaluated five SD datasets derived from ice, cloud, and land elevation satellite-2 (ICESat-2), passive microwave products (AMSR2 and GlobSnow), and reanalysis products (ERA5 and modern-era retrospective analysis for research and applications, version 2) across 13 representative snow-covered regions in the Northern Hemisphere. A novel dynamic upscaling approach was developed by integrating simple averaging with regression kriging for ICESat-2 SD data. Spatial matching of multisource datasets was then conducted using 3342 SD observation sites to comprehensively evaluate the uncertainties among the five datasets. The results indicate that the retrieval errors of each dataset are positively correlated with the mean SD across different regions. During snow accumulation and melt periods, ICESat-2 demonstrates significant advantages in nonforested areas with SDs ranging from 5 to 45 cm. Both passive microwave and reanalysis SD products demonstrate reliable performance during stable snow periods. However, products often miss snow during melt seasons despite ground confirmation. The causes of this phenomenon differ between the two datasets: passive microwave retrievals are primarily dominated by the physical properties of liquid water, whereas reanalysis products face limitations due to model structure and insufficient input data. In conclusion, the integration of station and ICESat-2 SD in nonforested regions may provide new possibilities for validating products.

Monitoring the Greenland Ice Sheet (GrIS) melt is critical for understanding cryospheric change and global sea-level rise. Existing passive microwave methods are limited by coarse resolution and detection biases. This study develops a novel multi-index decision-tree method for GrIS melt detection. Applied to enhanced-resolution brightness temperature data (up to 3.125 km), this method integrates dynamic thresholding of diurnal amplitude variation (DAV) from 37 GHz vertical polarization and Gaussian Mixture Model (GMM) classification from 19 GHz horizontal polarization, enabling precise identification of diverse melt types, including sporadic/intermittent melt, persistent melt, and supraglacial lake formation. Compared to traditional methods (245 K and MEMLS), our method significantly improves accuracy, achieving a 95.81% overall accuracy. For high-melt sites, 76.82% of AWS-recorded melt days are correctly classified. Long-term analysis (1988-2023) reveals a significant lengthening of the GrIS melt period: melt onset advanced by 0.31 d/yr, melt end delayed by 0.50 d/yr, and annual melt days increased by 0.35 d/yr (p < 0.01). These trends correlate with the negative summer North Atlantic Oscillation (r = -0.62) and increased Greenland blocking (r = 0.71, p < 0.01). This melt detection framework provides an observational basis for investigating ice sheet surface processes and supports more reliable assessments of melt-related impacts under climate change.

2025年1月2日中国科学院西北生态环境资源研究院 甘肃省遥感重点实验室的郝晓华团队和太原理工大学 矿业工程学院的赵子胜团队在《遥感学报》发文,介绍了其在高时空分辨率雪深数据获取领域的研究进展,建立了耦合深度学习模型与积雪微波辐射传输模型的降尺度雪深反演算法,为获取区域降尺度雪深产品提供保障。

Snow cover possesses high albedo and thermal insulation properties, greatly influencing climate change and responding rapidly to climatic variation, particularly on the Tibetan Plateau. The region's combination of low latitude and intense solar radiation further amplifies the effects of climate change on snow cover. During the spring and summer seasons, changes to snow cover patterns driven by climate change initiate shifts in hydrological cycles, affecting freshwater availability and vegetation growth. However, previous studies investigating snow cover changes across the Tibetan Plateau have yielded inconsistent results due to the use of diverse datasets across varying periods. This study aims to analyze the response patterns of snow cover to climate change on the Tibetan Plateau by comparing trends in snow cover duration (SCD) and snow depth (SD) during spring and summer from 1990 to 2019. Based on the spatial distribution of changes in these two parameters, we categorized regions into four distinct types: areas where both snow cover duration and snow depth increase (SCD+SD+), areas where snow cover duration increase while snow depth decreases (SCD+SD-), areas where snow cover duration decrease while snow depth increases (SCD-SD+), and areas where both snow cover duration and snow depth decreases (SCD-SD-). Rising temperatures are driving a shift from snowfall to rainfall in lower-altitude regions, while high-altitude areas remain predominantly below freezing, limiting this conversion. By analyzing temperature and precipitation changes in the four areas, we observed that the response of snow cover duration to global warming is limited at altitudes above 4700 m during spring and summer, with an increases in spring snowfall dominating the overall rise of snow cover duration. In contrast, the declines in snow cover duration observed in lower-altitude areas are primarily attributed to warming effects, regardless of variations in snowfall. The snow depth variation is mainly influenced by winter snowfall trends during spring and summer. Furthermore, the negative trends (p < 0.05) in summer snowfall are mainly influenced by the increase (p < 0.05) summer temperatures. The study aids in understanding the complexity and inconsistencies of snow cover changes under the combined effects of climate change and altitudes.

The snow density is a fundamental variable of the snow physical evolution processes, which can reflect the snowpack condition due to the thermal and gravitational compaction. Snow density is a bridge to transfer snow depth to snow water equivalent (SWE) for the snow water resources research. Therefore, it is important to understand the spatiotemporal distribution of snow density for the appropriate estimation of SWE. In this study, in situ snow densities from more than 6,000 stations in the Northern Hemisphere were used to analyze the spatial and temporal variations in snow density. The results displayed that snow density varied spatially and temporally in the Northern Hemisphere, with range of below 0.1 to over 0.4 g/cm3. The average snow densities in the mountainous regions of western North America, southeastern Canada, and Europe range from approximately 0.24 to 0.26 g/cm3, which is significantly greater than the values of 0.16-0.17 g/cm3 observed in Siberia, central Canada, the Great Plains of the United States, and China. The seasonal growth rates also present large spatial heterogeneity. The rates are over 0.024 g/cm3 per month in Southeastern Canada, the west mountain of North America and Europe, approximately 0.017 g/cm3 per month in Siberia, much larger than approximately 0.004 g/cm3 per month in other regions. Snow cover duration is a critical factor to determine the snow density. This study endorses the small snow density in China based on meteorological station observations, which results from that the meteorological stations are dominantly distributed in plain areas with relative short snow cover duration and shallow snow.

Vegetation change is an important indicator of ecological stability. The Heihe River Basin (HHRB), Shiyang River Basin (SYRB) and Shule River Basin (SLRB) in the Hexi Corridor of China have an arid climate and are ecologically fragile. Understanding vegetation changes and driving factors can guide the implementation of ecological restoration measures. The spatiotemporal characteristics of vegetation in the inland river basins of the Hexi Corridor during 1987-2023 were analyzed based on the NDVI using Landsat data, and the impacts of the driving factors on vegetation changes and spatial distributions were explored using residual analyses and Geodetector model. NDVI showed a spatial distribution pattern of high in the upper reaches and oases and low in the lower reaches. The spatial distribution of vegetation in the upper, middle, and lower reaches is influenced by different factors. Vegetation is most influenced by temperature and precipitation in the upper reaches of the basins, by population density in the middle reaches, and by a combination of natural factors and human activities in the lower reaches. NDVI trends indicate increases in the upper and middle reaches of the basins and the lower reaches of the HHRB, while the lower reaches of the SYRB and SLRB show decreases. The increasing trend is mainly driven by climate change, whereas the decrease is primarily due to human activities, climate change, or a combination of both. These findings are significant for vegetation growth in the Hexi Corridor's inland river basins and for the ecological management and sustainable development of areas.

Accurate snow depth (SD) data are essential for understanding and simulating hydrological processes, particularly in regions with uneven snow distribution and complex terrain, such as the Third Pole (TP). However, existing SD products often suffer from low spatial resolution, introducing uncertainties in regional snowpack and runoff modeling. In this study, we developed an integrated downscaling framework that combines deep learning (DL) with physical constraints to improve the resolution and accuracy of SD data for the TP region. First, a residual network incorporating multifactor spatial-terrain relationships is used to generate preliminary 500-m DL-downscaled SD data. Subsequently, by incorporating both snow cover fraction (SCF) data and snow depletion curve (SDC) as dual physical constraints, the new algorithm significantly improves the estimation accuracy in shallow snow areas. The proposed algorithm significantly enhances SD estimation accuracy, reducing the root-mean-square error (RMSE) from 1.99 to 0.75 cm, which represents a 62% improvement over the purely DL-based downscaling algorithm (RMSE = 1.02 cm). Notably, the improved algorithm demonstrates enhanced capability in terrain feature representation while simultaneously minimizing misclassification in snow-free areas and reducing overestimation in shallow snow regions. The results demonstrate progressive improvement in SD estimation accuracy across diverse geographical environments through stepwise validation of each downscaling step, confirming that integrating DL-based downscaling with physical constraints yields complementary advantages. Although the new algorithm significantly improves the accuracy of SD data, it is still underestimated in complex terrain areas and areas with SD exceeding 7 cm. Future research needs to deeply integrate physical constraint methods, comprehensively consider physical processes such as snow density, water content changes, and energy balance, and establish a more complete downscaling model.

Snowmelt water is a crucial freshwater resource in mid-to-high-latitude mountain areas. Snow depth is a key parameter in estimating snowmelt water. Current snow depth datasets have low spatial resolution and insufficient observational data for validating snow depth inversion in high-altitude areas. Based on ICESat-2 data, this study constructed a snow depth estimation model considering terrain and photon geometric relationships, obtaining a seasonal snow depth dataset for the Northern Hemisphere from 2018 to 2020 in flat areas (slope <= 5 degrees). Results indicate that ICESat-2 is effective for estimating snow depth in areas with gentle slopes, sparse vegetation, deep snowpacks, varying altitudes, and non-forest regions. (1) The ICESat-2 dataset shows 60.5% of biases within +/- 5 cm, with an average RMSE of 20.2 cm. (2) The best estimation comes from bare ground and herbaceous vegetation, with R-2 values of estimated and site observed values reaching 0.88 and 0.79, respectively. (3) The dataset tends to overestimate shallow snow and underestimate deep snow. This dataset offers the possibility of snow depth validation in high-elevation, flat, and sparsely vegetated regions.

Estimating snow parameters (e.g., snow cover, snow depth) at hillslope scales (<100 m) is an urgent but highly challenging research task. Remote sensing has become an indispensable tool for monitoring large-scale snow cover. However, existing studies on spatial patterns of snow cover typically focus on scales of 500-5000 m due to the trade-offs between temporal and spatial resolution in remote sensing sensors. This limitation hinders accurately representing the high spatial heterogeneity in mountain snow. To this end, we used a straightforward metric and developed an innovative 30 m average monthly Snow Cover Frequency (SCOF) map for the Tibetan Plateau (TP) utilizing high-resolution images from Landsat and Sentinel-2 satellites, spanning from 2000 to 2024. These maps provide a novel perspective for analyzing seasonal variations in snow cover and its relationship with topography. First, we designed a specific snow mapping algorithm tailored to shaded and low-illumination areas, which were depicted through terrain modeling. Validation against in situ observations and very-high-resolution remote sensing data demonstrated that snow could be effectively extracted even in challenging shadowed conditions. Next, SCOF maps were generated using snow cover data extracted from over 500,000 Landsat and Sentinel-2 images. Comparative analysis demonstrated that SCOF maps accurately characterize the spatial heterogeneity of mountain snow, with strong correlations with in situ observations, providing significantly enhanced spatial details compared to MODIS-derived SCOF maps. Finally, the spatial patterns of SCOF, along with seasonal variations and their relationship with topography, were meticulously documented. Key findings include: (1) SCOF exhibits apparent seasonal variations at elevations between 1500 and 6300 m, with the highest value typically observed in February and the lowest in August. However, above 7100 m, the highest SCOF occurs in July and the lowest in February, presenting an almost opposite seasonal pattern. Notably, although SCOF remains high at elevations of 6300-7100 m, seasonal variation are minimal, with the snow remaining relatively balanced across the seasons. (2) Generally, SCOF increases with elevation below 6300 m, showing a gradual rise below 5000 m but becoming significantly more rapid above this elevation. Above 7100 m, an intriguing phenomenon is observed: SCOF is unexpectedly lower in winter than in summer. In this region, SCOF in the warm season remains constant with increasing elevation, with slight increases in individual months, whereas in the cold season, SCOF shows a decreasing trend. (3) Elevation, distance to the coastline, and the topographic relief index are found to be important factors influencing snow distribution on the TP. The high-resolution SCOF maps presented in this study enhance our understanding of hillslope-level snow cover patterns in alpine regions lacking in situ observations, contributing to improved research on snow hydrology.

Passive microwave remote sensing has witnessed unprecedented progress in soil moisture (SM) estimation over the decades. However, it is challenging to estimate SM accurately due to the insufficient understanding of microwave emission mechanisms. A ground-based radiometry experiment is undertaken over an agricultural field, toward the reexamination and improvement of the microwave emission models and retrieval algorithms of SM and vegetation water content (VWC). This article reports the preliminary analysis of the experimental data and the calibration of the tau - omega model against the collected measurements. First, the collected multifrequency dual-polarized brightness temperature (TB) reflects the temporal variation of surface SM and a significantly negative correlation between them is observed, with the coefficient of determination ( R-2 ) and slope (S) of a linear fitting line ranging from 0.036 to 0.367 and from -16.7 to -81.5, respectively. Surface roughness and VWC impact the relationship between TB and SM, with variable R-2 and S observed. Second, the calibrated parameters have improved the model performance, with R-2 greater than 0.65 and root mean standard error (RMSE) less than 4.7 K at all frequencies and polarizations. The parameter values are frequency- and polarization-dependent, and the best performance of the model simulation is observed at V-polarization of L- and Ku-bands, with R-2=0.80 and RMSE =4.69 K at the L-band and R-2=0.74 and RMSE =2.9 K at the Ku-band. Overall, the experiment has provided valuable datasets for calibrating forward models and the calibrated model will facilitate the improvement of surface parameters (e.g., SM and VWC) retrieval.

Advancements in snow depth retrieval from passive microwave remote sensing data have mainly focused on influence of varying snow characteristics and forests, while neglecting the complicated mountainous terrain. Therefore, examining the influence of mountainous terrain on microwave radiation transmission of snowpack is beneficial for improvement of snow depth retrieval algorithms in mountainous areas. In this article, we established a microwave emission transfer model of snowpack in Mountainous areas within the framework of microwave emission transfer model of layered snowpack (MEMLS), thereafter, called MEMLS-T. MEMLS-T considers the influence of complicated terrain on the microwave radiation transmission of snowpack from three perspectives: the varied hill slopes alter the local incidence angle; the diverse hill slopes and aspects induce the polarization rotation; and the reduced sky visibility in mountainous regions results in an escalation of downward background radiation reaching the snow surface, as a consequence of the illumination from neighboring slopes. We simulate brightness temperatures at varying sky visibilities, slopes and aspects using MEMLS-T, and find that, in most conditions, the mountainous terrain weakens the brightness temperature difference between K and Ka bands (TbD), yet it also augment the TbD. The proportion of augmentation is approximately 8.9% for vertical polarization and 4.6% for horizontal polarization, and decreases as snow depth (or scattering intensity) increase or sky visibility decline. The uncertainties in snow depth retrieval in mountainous areas increase as snow depth increases. The brightness temperatures are simulated based on various spatial resolutions (30 m, 90 m, 270 m, 810 m) of the digital elevation model (DEM) of Qilian mountain and integrated into a grid of 6.25 km x 6.25 km. The results reveal that the coarser DEM results in less variation in brightness temperatures and TbD. Although the scale of DEM has a significant impact on the simulation of brightness temperature and TbD, its influence on average TbD at satellite scale is greatly reduced, resulting in approximately 3% variation in retrieved snow depth when using snow depth retrieval algorithm developed in flat terrain.

The single-channel algorithm (SCA), dual-channel algorithm (DCA), and multichannel algorithm (MCA) are commonly utilized in passive microwave remote sensing for estimating soil moisture (SM) and vegetation water content (VWC). However, determining the most effective algorithm(s) remains a challenge as there is a lack of comparative research on these algorithms. To this end, a comprehensive evaluation of these algorithms is presented using multifrequency radiometry data collected during a ground-based experiment. SM and VWC are separately estimated through a single-objective estimation (SOE) or simultaneously estimated through a dual-objective estimation (DOE) framework. The study reveals that SOE yields better results than the DOE framework in estimating SM and VWC using the SCA, DCA, and MCA algorithms, with decreasing accuracy performance as follows: MCA-SOE>DCA-SOE>SCA-SOE>MCA-DOE>SCA-DOE>DCA-DOE. Within the MCA-SOE scheme, the utilization of five-channel TBs (i.e., combination of V-polarized TB at L-band, dual-polarized TBs at Ku- and Ka-bands) leads to the most accurate SM and VWC retrievals, with an R value of 0.824 and a root mean square error (RMSE) (unbiased RMSE, ubRMSE) of 0.039 (0.037) m(3)/m(3) for SM, and an R value of 0.771 and an RMSE (ubRMSE) of 0.707 (0.705) kg/m(2) for VWC. An exceptional case shows that SM retrieval from a single Ku-band TB slightly exceeds that from MCA-SOE in $R$ value (0.831 versus 0.824), due to the less sensitive Ka-band TBs introduced in MCA-SOE. This evaluation underscores the benefits of employing SCA, DCA, and MCA algorithms for SM and VWC estimation, providing insights for enhancing SM retrieval methodologies and selecting optimal frequency and polarization.

Passive microwave (PMW) remote sensing has emerged as a promising approach for the development of long-term, global, and daily fractional snow cover (FSC) products. Despite its potential for FSC retrieval, the method remains at an early stage of development with limited understanding of the underlying mechanisms and insufficient research on its applicability and reliability. This study systematically assessed the potential of PMW remote sensing utilizing AMSR2 brightness temperature (Tb) data in Western China (WCN) during the snow-stable period. The proposed model integrates PMW observations with auxiliary information, employing machine learning techniques capable of capturing nonlinear relationships between predictors and FSC. Key variables were selected using the simulated annealing algorithm (SA) and recursive feature elimination algorithm (RFE) to improve model performance and multiple machine learning methods,viz. Linear Regression, Tree Methods, Support Vector Machine, Ensemble Algorithms, Nearest Neighbor Regression, MARS, and NNET were constructed. Results demonstrated that PMW showed considerable potential and could be effectively applied toFSC retrieval in WCN. Ensemble algorithms, such as RF, LightGBM, and BART, generally demonstrated stronger performance than the other methods (R2 = 0.762, RMSE = 0.191, MAE = 0.121), and the RFE-LightGBM model yielded satisfactory overall performance (R2 = 0.792, RMSE = 0.182, MAE = 0.111). The results were validated against in situ observations and the MODIS FSC products, showing excellent agreement and reliable retrieval accuracy with each dataset.Error analysis revealed systematic overestimation at low FSC values (<0.2) and underestimation at high FSC values (>0.8). The results also confirmed strong temporal transferability across varying training and testing periods. This research supports the use of passive microwave data for long-term snow cover monitoring since the 1980s, and highlights its relevance to cryospheric and climate research.

Snow cover is one of the most important factors controlling Arctic ecosystems' microclimate and plant growth conditions in Arctic ecosystems. Climate change has impacted the timing and spatial variability of both snow cover, and worldwide vegetation phenology across the globe. However, the mechanisms by which snowpack factors regulate the onset of the growing season remain to be thoroughly investigated, particularly under varying climatic conditions and growth stages. In this study, we investigated the influence of snow characteristics on vegetation phenology across different growth stages. Specifically, we analyzed the spatiotemporal dynamics of vegetation and snow phenology in Alaska from 2001 to 2021, assessed the partial correlations between key phenological groups under controlled temperature and precipitation conditions, and quantified the contributions of climatic variables and snow cover to vegetation phenology across various growth-cycle phases. The results revealed that grassland and forest phenology responded strongly to variations in snowmelt timing (r> 0.5, p < 0.05). In contrast, although phenological responses in wetlands were also statistically significant (p < 0.05), the average correlation was weaker (mean r z 0.45). Temperature was found to be the primary driver of vegetation phenology change during the peak growth periods, whereas snow temperature and depth were crucial drivers during the transitional growth phases. From 2001 to 2021, the changes in vegetation phenology in Alaska were more pronounced than those in snow dynamics. Notwithstanding the significant role of other co-varying drivers of vegetation-phenological shifts, the influence of snow phenology was crucial. This study elucidates the role of snowpack phenology in regulating vegetation dynamics under changing climatic conditions and growth cycles.