Reconciling historical changes in the hydrological cycle over land

Anna M. Ukkola,Martin De Kauwe,Sanaa Hobeichi,Gab Abramowitz,Hylke Beck,Jason P. Evans,Andy Pitman

doi:10.1038/s41612-022-00240-y

Abstract

The sixth Intergovernmental Panel on Climate Change (IPCC) assessment report confirms that global warming drives widespread changes in the global terrestrial hydrological cycle, and that changes are regionally diverse. However, reported trends and changes in the hydrological cycle suffer from significant inconsistencies. This is associated with the lack of a rigorous observationally-based assessment of simultaneous trends in the different components of the hydrological cycle. Here, we reconcile these different estimates of historical changes by simultaneously analysing trends in all the major components of the hydrological cycle, coupled with vegetation greenness for the period 1980–2012. We use observationally constrained, conserving estimates of the closure of the hydrological cycle, combined with a data assimilation approach and observationally-driven uncertainty estimates. We find robust changes in the hydrological cycle across more than 50% of the land area, with evapotranspiration (ET) changing the most and precipitation (P) the least. We find many instances of unambiguous trends in ET and runoff (Q) without robust trends in P, a result broadly consistent with a “wet gets wetter, but dry does not get drier”. These findings provide important opportunities for water resources management and climate risk assessment over a significant fraction of the land surface where hydrological trends have previously been uncertain.

Highlights

Understanding how the future hydrological cycle will change at regional scales is crucial for impact assessment and adaptation planning[1]
We find that 51% of the land surface has experienced robust changes in at least one component of the hydrological cycle
Robust changes in the hydrological cycle are observed over more than 50% of the global land area even when robust measures for assessing the significance of trends are implemented

Summary

INTRODUCTION

Understanding how the future hydrological cycle will change at regional scales is crucial for impact assessment and adaptation planning[1]. Some studies have suggested that hydrological cycle trends are consistent with ‘wet gets wetter and dry gets drier’[9,16], while others have found that both this and the opposite pattern ‘dry gets wetter and wet gets drier’ is detectable on land[3,17]. These conflicting findings may stem from different interpretations of ‘drying’, ‘aridity’ and ‘intensification’, depending on the metrics and spatial-scale aggregation methods used[4,18]. The analysis of trends utilises multiple products and observational sources in the derivation of each budget component, as well as uncertainty bounds informed by observational constraints. Changes in the hydrological cycle are described at the grid scale (0.25°) and over precipitation regimes derived by clustering the land based on precipitation mean and variability (Supplementary Fig. 1 and Supplementary Table 2)

RESULTS

DISCUSSION

METHODS

CODE AVAILABILITY