Spin‐up is essential to provide initial conditions for land surface models (LSM) when they cannot be given reliably as in the application to regional permafrost change studies. In this study, the impacts of spin‐up strategy including total spin‐up length and cycling scheme on modeling of permafrost dynamics on the Qinghai‐Tibet Plateau (QTP) were evaluated through two groups of experiments using a modified Noah LSM. The first group aims to test different total spin‐up lengths and the second group for different cycling schemes. The results show that the presence of permafrost prolongs the convergence of the model. Vertically, the slowest convergence is observed at the permafrost table. The insufficiency of total spin‐up length is prone to underestimate permafrost area and overestimate the degradation rate. Different cycling schemes considerably affect the resulting initial thermal fields and result in degradation rates with a difference of 3.37 × 10³ km²/a on the QTP, which exceeds the difference (2.92 × 10³ km²/a) in the degradation rates reported in existing studies. The multi‐year cycling scheme is generally preferred, but overlong cycle length should be avoided to prevent the introduction of climate change trends in the spin‐up period. We recommend a spin‐up strategy of a 500‐year cycling with the first 5‐ to 10‐year of forcing for modeling permafrost on the QTP with the Noah LSM. Our findings highlight the importance of the spin‐up strategy, which is usually neglected in present LSM‐based permafrost modeling studies.

The migration of rivers in permafrost landscapes has critical implications for riverine infrastructure, ecosystem stability, and carbon cycling, yet its magnitude and underlying mechanisms remain poorly understood. Here, we leverage four decadal satellite imagery, hydrological observations, and permafrost modeling to investigate meander migration dynamics on the Tibetan Plateau. Our data show that the migration rates of permafrost rivers have increased by 34.6% from 1987 to 2022, in response to the combined effects of increased discharge, riverbank destabilization driven by ground ice melt and extended thawing days (increased by 35 days). In contrast, rivers flowing across seasonally frozen ground exhibited a decline in migration rate by 11.1%, driven by vegetation greening and riverbank stabilization. In a future warming climate for the Tibetan Plateau, the migration rates of permafrost rivers are anticipated to further accelerate, potentially threatening riverine infrastructure safety and aquatic ecosystems, and intensifying the permafrost carbon cycle.

River‐controlled permafrost dynamics are crucial for sediment transport, infrastructure stability, and carbon cycle, yet are not well understood under climate change. Leveraging remotely sensed datasets, in‐situ hydrological observations, and physics‐based models, we reveal overall warming and widening rivers across the Tibetan Plateau in recent decades, driving accelerated sub‐river permafrost thaw. River temperature of a representative section (Tuotuohe River) on the central Tibetan Plateau, has increased notably (0.39°C/decade) from 1985 to 2017, facilitating heat transfer into the underlying permafrost via both convection and conduction. Consequently, the permafrost beneath rivers warms faster (0.37°C–0.66°C/decade) and has a ∼0.5 m thicker active layer than non‐inundated permafrost (0.17°C–0.49°C/decade). With increasing river discharge, the inundated area expands laterally along the riverbed (16.4 m/decade), further accelerating permafrost thaw for previously non‐inundated bars. Under future warmer and wetter climate, the anticipated intensification of sub‐river permafrost degradation will pose risks to riverine infrastructure and amplify permafrost carbon release.

Rivers regulate land-ocean total phosphorus (TP) fluxes critical to ecosystem health and food security, yet global dynamics remain poorly understood due to limited observations. Here, we develop a machine learning framework integrating multimodal data and 280,000 TP measurements to reconstruct TP flux patterns over 1980–2019 across 420 major rivers. Results reveal a deceptive global equilibrium. While TP flux declines in the Northern Hemisphere, driven by dam trapping, particularly in Western Europe (−16.2%) and Eastern Asia (−8.7%), it rises in the Southern Hemisphere due to increased fertilizer use and deforestation, especially in Southern Africa (+15%) and the Malay Archipelago (+20.3%). Notably, the number of small rivers with rising TP flux is nearly double that of large rivers. This growing TP export from small rivers and Southern Hemisphere basins may intensify eutrophication, expand hypoxic zones, and threaten fishery yields. Our findings highlight a shifting global phosphorus landscape and underscore the need for more targeted, sustainable phosphorus management strategies.

Frozen ground and precipitation seasonality may strongly affect hydrological processes in a cold alpine basin, but the calibration of a hydrological model rarely considers their impacts on model parameters, likely leading to considerable simulation biases. In this study, we conducted a case study in a typical alpine catchment, the Babao River basin, in Northwest China, using the distributed hydrology–soil–vegetation model (DHSVM), to investigate the impacts of frozen ground type and precipitation seasonality on model parameters. The sensitivity analysis identified seven sensitive parameters in the DHSVM, amid which soil model parameters are found sensitive to the frozen ground type and land cover/vegetation parameters sensitive to dry and wet seasons. A stratified calibration approach that considers the impacts on model parameters of frozen soil types and seasons was then proposed and implemented by the particle swarm optimization method. The results show that the proposed calibration approach can obviously improve simulation accuracy in modeling streamflow in the study basin. The seasonally stratified calibration has an advantage in controlling evapotranspiration and surface flow in rainy periods, while the spatially stratified calibration considering frozen soil type enhances the simulation of base flow. In a typical cold alpine area without sufficient measured parametric values, this approach can outperform conventional calibration approaches in providing more robust parameter values. The underestimation in the April streamflow also highlights the importance of improved physics in a hydrological model, without which the model calibration cannot fully compensate the gap.