The terrestrial water cycle (TWC) is a critical component of the climate system, significantly impacting water resources, agriculture, and ecosystems globally. Global warming has demonstrably strengthened the TWC, leading to increased global annual total precipitation (PRCPTOT). However, the spatial distribution of these changes is heterogeneous, with differing trends and magnitudes across climatic regions. While the conventional expectation is that dry areas will become drier and wet areas wetter, studies have observed a faster and more substantial increase in PRCPTOT in dry regions compared to wet regions. This increase, however, is often offset or overcompensated by warming-induced increases in evapotranspiration, leading to variable impacts on water availability despite heightened risks of short-duration rainfall extremes. Understanding the varying responses of the TWC to warming in different regions is crucial for effective infrastructural planning, water resource management, and sustainable development. Anthropogenic climate change (ACC) intensifies the TWC through atmospheric thermodynamic and dynamic processes; warming increases atmospheric humidity and modifies circulation patterns, leading to regional upward motions and moisture convergence, which in turn enhance PRCPTOT. The varying rates of warming and precipitation response per unit warming between dry and wet regions, however, remain unclear. Different external forcings (e.g., greenhouse gas emissions (GHG) and anthropogenic aerosol emissions (AER)) have contrasting effects on warming, with GHG leading to global warming and AER causing regional cooling. The combined effects of these forcings could explain the divergent changes in PRCPTOT observed in recent decades. Future projections of these changes are also uncertain due to the anticipated increase in GHG and decrease in AER. Prior research has quantified the impact of ACC on PRCPTOT in various global regions, finding a high likelihood of ACC explaining observed increases in the Northern Hemisphere. However, the significantly greater increase in PRCPTOT in dry compared to wet regions necessitates a more comprehensive assessment of the ACC signal across both climate types. This study aims to address this gap by evaluating whether the human-induced intensification of the global TWC is primarily driven by the increase in PRCPTOT over dry regions.

Existing literature highlights the strengthening of the global terrestrial water cycle (TWC) due to global warming, evidenced by increased global annual total precipitation (PRCPTOT). However, considerable spatial heterogeneity exists in these trends, with varying magnitudes and directions across different climatic regions. Contrary to the widely held belief that dry regions will experience further drying and wet regions further wetting, recent studies indicate a faster and stronger increase in PRCPTOT in dry regions compared to wet regions. This seemingly counterintuitive observation has been linked to the offsetting or overcompensating effects of warming-induced evapotranspiration increases. Although PRCPTOT increases in dry regions, water availability may not necessarily increase, and the risk of short-duration rainfall extremes grows. The literature also points to the role of anthropogenic climate change (ACC) in intensifying the TWC through both thermodynamic and dynamic atmospheric processes. Warming enhances atmospheric humidity and modifies circulation patterns, leading to moisture convergence and increased PRCPTOT. The differential responses of dry and wet regions to these processes, however, require further investigation. The influence of various external forcings, like greenhouse gas emissions (GHG) leading to global warming and anthropogenic aerosol emissions (AER) causing regional cooling, complicates this picture. The combined effects of GHG and AER could explain the observed divergent changes in PRCPTOT across regions. Detection and attribution studies have quantified ACC's influence on PRCPTOT in different regions. However, the disproportionate increase in PRCPTOT in dry regions necessitates further analysis to determine if the ACC signal is equally detectable in both dry and wet climates.

This study investigates changes in normalized PRCPTOT over dry and wet land regions using multiple observed datasets (HadEX3, CRU, and GPCC) and climate simulations from the Coupled Model Intercomparison Project Phase 6 (CMIP6). The observed datasets are gauge-based and cover a substantial portion of the global land surface (excluding Antarctica), with data gap information provided by HadEX3. Grid cells with less than 10% missing data were selected for analysis. Dry and wet regions were identified based on the climatological mean PRCPTOT, with the bottom and top 30% of the grid cells defining these regions respectively. This methodology was tested for sensitivity by varying the threshold percentages (20% and 40%). The CMIP6 simulations under historical (ALL, NAT, ANT, GHG, AER) and Shared Socioeconomic Pathway (SSP) scenarios (SSP126, SSP245, SSP370, SSP585) were used to confirm the observed differences in PRCPTOT changes and project future trends. The precipitation-temperature relationship and the response of atmospheric moisture to global warming were investigated to explain the observed differences. Two detection and attribution methods were employed: a pattern-based method using rotated empirical orthogonal functions (REOF) to extract fingerprints of ACC from CMIP6 simulations, and an optimal fingerprinting method to quantify the contributions of different external forcings to observed PRCPTOT changes. The REOF method involves projecting observed and simulated PRCPTOT onto model-based fingerprints to quantify the ACC signal and its significance, while the optimal fingerprinting method regresses observed against simulated PRCPTOT to estimate scaling factors and quantify the contributions of different forcings. Moisture content and transport were analyzed using vertically integrated water vapor (VIWV) and vertically integrated water vapor transport (IWVT), respectively, calculated from CMIP6 data and normalized for consistency. The sensitivity of results to the choice of thresholds used to identify dry and wet regions was also examined. Finally, precipitation minus evapotranspiration (PME) was used in place of PRCPTOT to determine if the use of PRCPTOT alone was introducing significant bias.

The study's key findings demonstrate a significant intensification of the terrestrial water cycle (TWC) in dry regions compared to wet regions. Analysis of three gridded observed datasets (HadEX3, CRU, GPCC) and CMIP6 simulations revealed consistently faster increases in PRCPTOT in dry regions (1.1% per decade in observations, 1.2% in ALL simulations) compared to wet regions (0.2% per decade in both observations and ALL simulations) from 1961 to 2018. This difference is substantially amplified when considering only anthropogenic forcing (ANT), showing a tenfold difference in the rate of increase. Simulations under different scenarios (GHG, AER, NAT) highlighted the crucial role of anthropogenic forcings, particularly GHG, in driving this disproportionate increase in PRCPTOT in dry regions, while AER contributes to a decrease in PRCPTOT in wet regions. Future projections under various SSPs indicate that this trend of faster precipitation increases in dry regions is projected to persist and intensify with higher emission scenarios. The faster warming rates in dry regions (0.33 K decade⁻¹ compared to 0.23 K decade⁻¹ in wet regions under ALL forcing) and a greater PRCPTOT response to warming (3.07% K⁻¹ in dry regions versus 0.98% K⁻¹ in wet regions) explain this phenomenon. Analysis of moisture content (VIWV) and moisture transport (IWVT) showed no significant difference in the rate of increase in moisture content per unit warming between the two regions, suggesting thermodynamic effects are not the primary drivers of the observed PRCPTOT difference. However, a marked difference was observed in moisture transport, with significantly stronger increases in dry regions (7.7–11.1% per unit warming under SSP245/SSP585) than in wet regions (3.6–3.8% per unit warming). This suggests that dynamic effects, amplified by land-atmosphere feedbacks, contribute significantly to the enhanced TWC in dry regions. Detection and attribution analyses using a pattern-based REOF method and an optimal fingerprinting method provided further quantitative evidence for ACC's role. The REOF analysis revealed a detectable ACC signal (SNR >1.64) in dry and dry+wet regions but not in wet regions alone. The optimal fingerprinting method confirmed these results, showing that the simulated responses under ALL, ANT, and GHG forcings are only detectable in observed PRCPTOT in dry regions. Quantifying the contributions of external forcings indicated that 5.63–7.39% (2.44–2.80%) of the observed PRCPTOT increase in dry (wet) regions can be attributed to ACC (based on OLS estimations). The results were not sensitive to the choice of thresholds for defining dry and wet regions, nor to the inclusion or exclusion of evapotranspiration in the analysis.

The study's findings robustly demonstrate an intensified TWC in dry regions, predominantly attributable to ACC. The faster warming rates and higher PRCPTOT response rates in dry regions, driven by increased moisture transport, are key factors. Thermodynamic effects (increased moisture content) play a less prominent role compared to the dynamic effects (enhanced moisture transport) of ACC, particularly GHG. The stronger moisture transport in dry regions, amplified by land-atmosphere feedbacks, stands out as a critical driver. In contrast, wet regions, often located in monsoon zones, show a less pronounced response. The weakening of the Walker and Hadley circulations under GHG-induced warming likely inhibits moisture transport to these regions. Anthropogenic aerosol emissions (AER) exert a complex influence, increasing PRCPTOT in dry regions through dynamic effects and decreasing it in wet regions through both thermodynamic and dynamic effects. The study's results highlight the importance of considering regional variations when assessing the impact of ACC on the global TWC. The use of PRCPTOT as a proxy for the TWC's intensification, while acknowledging limitations regarding other hydro-meteorological fluxes and stores, remains justified by the direct impacts of PRCPTOT on water resources, agriculture, and ecosystems. The finding that the global ACC signal in PRCPTOT predominantly originates from dry regions necessitates a reassessment of ACC signals in other TWC components.

This research demonstrates a significant human-induced intensification of the terrestrial water cycle in the world's dry regions. The faster warming rates and increased PRCPTOT response in dry regions are primarily attributed to enhanced moisture transport, rather than changes in moisture content. Detection and attribution analyses strongly support the role of anthropogenic climate change, particularly greenhouse gas emissions. These findings emphasize the importance of region-specific considerations when evaluating the impacts of climate change on the global water cycle and highlight the potential for both benefits (alleviating water scarcity) and risks (increased flooding) in dry regions. Future work should focus on further refining regional climate models, incorporating land-atmosphere feedbacks in greater detail, and investigating the socio-economic consequences of the intensified water cycle in drylands.

The study's limitations include the reliance on gridded precipitation datasets which can mask sub-grid variations and the use of PRCPTOT as a primary indicator for TWC intensification. While PME analysis was conducted to address the potential bias introduced by neglecting evapotranspiration, other hydro-meteorological factors like runoff, soil moisture and snow were not directly considered. The study primarily focuses on the effects of large-scale climate forcings and might not capture the influence of smaller scale processes or local anthropogenic activities that influence the water cycle. Furthermore, the uncertainty associated with climate model projections should be taken into account when interpreting the future scenarios.