Warming and drying over the central Himalaya caused by an amplification of local mountain circulation

The central Himalaya, a critical source of freshwater for hundreds of millions downstream, has undergone rapid warming and decreasing monsoon precipitation in recent decades, contributing to significant glacier retreat. Diagnosing the meteorological drivers is challenging due to sparse observations and limitations of reanalyses and global models in representing orographic diurnal cycles. Reanalyses indicate an upper-tropospheric circulation trend over Asia since 1979 involving enhanced westerlies and an anticyclonic anomaly linked to the Silk Road Pattern. The study hypothesizes that recent regional circulation changes have amplified the mountain–valley diurnal circulation over the Himalaya, leading to daytime warming and nocturnal drying, thereby influencing the observed hydroclimatic trends.

Prior work documents increasing temperatures and decreasing monsoon precipitation over the central Himalaya, with associated glacier retreat. Reanalyses reveal changes in upper-tropospheric circulation across Asia since the late 1970s, including a westerly enhancement and an anticyclonic trend over East Asia consistent with Silk Road Pattern variability. Contrasting glacier responses between the central Himalaya (retreat) and Karakoram (more heterogeneous trends) have been reported. The Himalayan diurnal cycle—daytime anabatic upslope winds promoting peak precipitation and nighttime katabatic downslope winds shifting precipitation toward foothills—is well established from observations and modeling. Previous evaluation of this WRF downscaling showed reasonable agreement with observed diurnal precipitation, temperature, and winds, supporting use for analyzing diurnal trends.

A 36-year dynamical downscaling was performed using the Advanced Research WRF (version 3.7) with nested domains at 20 km and 6.7 km grid spacing, forced by NCEP CFSR reanalysis for initial and boundary conditions. Simulations span April 1979–March 2015 (36 separate 13-month runs), with March as spin-up and analyses focusing on JJA. The inner domain covers most of the Tibetan Plateau, Himalaya, and Karakoram. No convective parameterization was used in the 6.7-km inner domain (gray-zone resolution), with the Noah-MP land surface model, Thompson microphysics, and RRTMG radiation schemes employed; temperature and other fields were spectrally nudged to the reanalysis. A subregion (86–92°E, 26–29°N) encompassing the central/eastern Himalaya glaciated areas was analyzed for JJA. Model output was archived 3-hourly, enabling diurnal-cycle analysis. For each variable, at each grid point and time of day, JJA means were computed annually to form 36-element time series. Trends and their significance were assessed via a two-tailed Monte Carlo permutation test (1000 shuffles), identifying grid points where the observed trend magnitude exceeded 95% of shuffled trends. Three-dimensional fields were interpolated to pressure levels before analysis. Precipitation trends correspond to 3-hour accumulations, while other variables were averaged over matching 3-hour windows. Additional line plots contrasted early (1979–1988) vs late (2005–2014) decades with Student’s t-tests for diurnal means. Daily-mean trends were also computed for context. Reanalysis comparisons (ERA-Interim, MERRA-2, JRA-55) assessed robustness of large-scale circulation trends.

• The diurnal cycle over the Himalaya is climatologically characterized by afternoon upslope (southerly) winds and peak precipitation over ridges, and nighttime downslope (northerly) winds with weaker, more evenly distributed precipitation and foothill maxima. • Since 1979, high-elevation (generally >3 km) daytime temperatures have increased, while nocturnal warming is weak and often not significant. • Near-surface winds exhibit significant diurnal asymmetry in trends: enhanced upslope (southerly) winds by day and stronger downslope (northerly) winds by night, with the nocturnal downslope trend being stronger and more widespread. Cross-section analyses show increased upslope acceleration between ~800–500 hPa during daytime and increased downslope acceleration between ~800–600 hPa at night. • Precipitation has decreased significantly at night, primarily from about 1800–0300 local time, and mainly at high elevations (>3 km). Low elevations show little to no nocturnal drying, consistent with enhanced valley convergence under stronger downslope flows. • Daytime high-elevation warming is linked primarily to large-scale tropospheric warming (ingested via reanalysis nudging), with no consistent positive trends in net shortwave or net longwave at the surface in early afternoon. Despite a modeled decrease in downward shortwave over peaks (due to increased afternoon cloudiness associated with enhanced upslope convergence), a concurrent decrease in surface albedo (reduced snow cover) increases absorbed shortwave, yielding near-neutral net shortwave trends. • Nighttime drying is associated with widespread reductions in upper- and mid-tropospheric cloud fraction near sunset, enhancing radiative cooling of mountain slopes under clearer skies. This increased radiative cooling coincides spatially with strengthened katabatic winds and reduced nocturnal precipitation at high elevations. • The regional upper-tropospheric circulation trend (anticyclonic center linked to the Silk Road Pattern) is present across multiple reanalyses, supporting the robustness of the large-scale driver downscaled by WRF.

The results support the hypothesis that regional circulation changes over Asia have amplified the orographic diurnal cycle in the central Himalaya, producing contrasting day–night hydroclimatic trends: daytime warming with stronger anabatic winds and nighttime drying with stronger katabatic winds. Enhanced daytime upslope flow promotes convergence and cloud over peaks but, coupled with reduced albedo from decreased snow cover, does not offset warming at high elevations. At night, clearer skies intensify radiative cooling on slopes, reinforcing downslope winds and reducing high-elevation precipitation. These meso-orographic processes help explain observed warming and drying patterns and likely contribute to rapid glacier retreat and altered water resources downstream. The findings also reveal limitations in GCMs and coarse regional models that under-represent orographic diurnal processes and nocturnal high-elevation precipitation, implying that current projections of Himalayan hydroclimate—especially precipitation—may be biased, with implications for glacier mass balance and runoff forecasts.

A multi-decadal high-resolution downscaling reveals that an amplification of the mountain–valley diurnal circulation over the central Himalaya under recent regional circulation trends has produced daytime high-elevation warming, enhanced upslope winds, and significant nocturnal drying associated with strengthened downslope winds and reduced cloudiness. These processes likely contribute to accelerated glacier mass loss and altered hydrological regimes. The study underscores the need for high-resolution, process-aware climate projections over complex terrain to capture orographic diurnal dynamics. Future work should: (1) extend and refine downscaling with convection-permitting resolutions and improved land–ice initialization; (2) better constrain radiative–albedo feedbacks with observations of snow/ice cover; (3) assess robustness across physics schemes and reanalyses; and (4) couple to glacio-hydrological models to quantify impacts on glacier mass balance and downstream water resources.

Key limitations include: (1) 6.7-km grid spacing in the inner domain lies in the convective gray zone and may not fully resolve convective–orographic systems; no convective parameterization was used in the inner domain. (2) Surface properties (e.g., snowpack, albedo) were initialized from reanalyses with a 3-month spin-up; the magnitude of declining snow cover during monsoon may be underestimated, potentially underestimating daytime warming and upslope wind trends. (3) Daytime cloud trends are more uncertain due to convective influences at this resolution; nocturnal cloud trends are considered more reliable. (4) The balance between reduced downward shortwave and reduced surface albedo leading to near-neutral net shortwave may be model-configuration dependent; thus net shortwave trend conclusions are tentative. (5) Trends ultimately depend on reanalysis forcing (CFSR), although similar large-scale patterns are present in ERA-Interim, MERRA-2, and JRA-55.