A global assessment of change in flood volume with surface air temperature
Introduction
The impact of climate change on the water cycle continues to be a topic of debate (Donat et al., 2016; Koutsoyiannis, 2020). It is generally expected that precipitation extremes will increase (Fowler et al., 2021), but across large parts of the world flood magnitudes have declined (Do et al., 2017). Meanwhile, on average mean rainfalls are predicted to increase but so is the occurrence of drought and hence water scarcity. Most increases in rainfalls have been observed over the tropics (Beck et al., 2019) with decreases elsewhere, while most warming is occurring over the land relative to the oceans, leading to climate change affecting different parts of the world disproportionally.
Rising temperatures will intensify heavy rainfall events (Fowler et al., 2021; Craig, 2010; Trenberth et al., 2003; Kim et al., 2022) as the atmospheric moisture‐holding capacity increases as per the Clausius‐Clapeyron (C‐C) relation (Westra et al., 2014; Magan et al., 2020). Due to the CC relationship there is a 6–7%/ °C increase in the saturation vapor pressure, creating a corresponding increase in the maximum amount of low-level moisture in the atmosphere (Held and Soden, 2006). Assuming constant relative humidity (Lenderink and Attema, 2015), this can be translated to a similar increase in extreme precipitation with higher temperatures (Sun et al., 2021), a measure that has been termed scaling or sensitivity (expressed in%/ °C units) in the literature (Trenberth et al., 2003).
As long-lasting or intense precipitation is often the main cause of pluvial flooding (Smith and Ward, 1998), such positive scaling estimates along with precipitation trends have been used to suggest future increases in flooding (Ivancic and Shaw, 2015; Seneviratne et al., 2012; Kundzewicz et al., 2014; Field et al., 2012a). This is despite the considerable uncertainty associated with future precipitation projections (Kim et al., 2020; Hosseinzadehtalaei et al., 2017), and little observational evidence to suggest that flood magnitudes have increased (Field et al., 2012b; Hirabayashi et al., 2013). In fact, observational records often present more evidence for a decrease in annual flood maxima (Do et al., 2017; Wasko and Nathan, 2019; Gudmundsson et al., 2019), despite increases in precipitation with climate change being well documented (Westra et al., 2014; Sun et al., 2021; Alexander et al., 2006).
For example, Shen and Chui (2021a) found that both scaling in extreme precipitation and streamflow across the Continental United States (CONUS) are not always on the same direction, depending on whether the extreme streamflow is more correspondent to extreme precipitation or other factors/processes like soil moisture and snowmelt. The inconsistency in both precipitation and streamflow scaling was also identified in South and Central Asia (Ghausi and Ghosh, 2020) where daily scaling of extreme precipitation becomes negative at high temperature while extreme streamflow scaling against temperature remains strongly positive. However, Shen and Chui (2021b) found that flood-temperature scaling is often similar to that of extreme precipitation-temperature from an analysis for three different hydroclimatic regions in the United States for short durations and impervious surfaces. Indeed, an evaluation of historical trends across Australia, where smaller floods have been decreasing and larger floods increasing found good agreement with peak-flow temperature sensitivities (Wasko, 2021).
In addition to flood control (Milly et al., 2008), water resource management has been argued to be under threat as temperatures rise worldwide (Karr, 1991; Lenzen and Foran, 2001). While some studies have found increases in streamflow globally (Labat et al., 2004; Probst and Tardy, 1987), an overall drying trend was found using the largest 925 rivers around the world (accounting for 73% of global runoff) by Dai et al. (2009), consistent with Milliman et al. (2008) who used a sample representing approximately half of the world's discharge. Indeed, higher annual temperatures have been linked to lower annual streamflows (Milly et al., 2018) and higher seasonal temperatures to greater water insecurity (Hettiarachchi et al., 2022).
Although Wasko and Sharma (2017) found negative peak flood scaling with temperature for frequent flood events, whether this translates into a reduction of water availability and water supply reliability, remains an open question. Given the importance of flood volume in ensuring water supply reliability in reservoir systems worldwide, this study builds off the above-mentioned negative flood scaling results to focus on flood volume and address the following questions: (1) Does there exist consistency in the scaling relationship for flood peak and flood volume? (2) If inconsistent, what factors can be correlated to the different pattern of change for flood volumes compared to peaks? and (3) What implications do these scaling relationships have on water security, both in the context of flood insecurity, and more importantly, given the focus on flood volumes, on water supply insecurity worldwide?
Section snippets
Data
Streamflow data from anthropogenically unaffected catchments from across the world, sourced from the Global Runoff Data Centre (GRDC) dataset (GRDC, 2015), are used in this study. The GRDC dataset provides daily streamflow for 6946 stations worldwide and has been successfully used in multiple global studies (Do et al., 2017; Lee et al., 2015; Milly et al., 2018; Wasko et al., 2020). The Global Historical Climatology Network (GHCN) dataset is used to provide daily air temperature and daily
Methods
We applied a three-step approach in sequence for calculating the flood peak and volume scaling for each catchment: (1) baseflow separation from an observed entire hydrograph; (2) identification of flood events from the remaining quickflow; (3) quantile regression of flood events with their corresponding observed temperature. Each step is detailed below.
Flood volume and flood peak scaling
We begin by presenting the 50th percentile volume and peak scaling at each station as the 50th percentile represents frequent flood events that are of interest here (Fig. 3). To aid visualisation, point estimates were spatially interpolated over the two-dimensional space using bilinear interpolation with the original point estimates and the spatially interpolated maps of scaling presented in Figs. S1–S3 in supporting information.
Mostly negative volume scaling is present when results are
Conclusions
In this study, we did not seek to find a causality between air temperature and flood runoff. Rather, based on the assumption that flood volume is representative of water resources augmentation in supply systems, this study investigated the flood volume-temperature sensitivity represented by scaling (α) at a global scale as a measure of water supply security. We identified three key findings in the framework of the flood volume-temperature sensitivity as follows.
First, in most regions, except
CRediT authorship contribution statement
Wei He: Validation, Formal analysis, Investigation, Data curation, Writing – original draft, Visualization. Seokhyeon Kim: Conceptualization, Methodology, Writing – review & editing, Project administration, Supervision. Conrad Wasko: Conceptualization, Methodology, Writing – review & editing, Funding acquisition. Ashish Sharma: Conceptualization, Methodology, Writing – review & editing, Supervision, Funding acquisition, Supervision.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
Conrad Wasko receives funding from the University of Melbourne McKenzie Postdoctoral Fellowships Program and ARC project DE210100479. This research was partially supported by a Discovery Project (DP200101326, Assessing Water Supply Security in a Nonstationary Environment) funded by the Australian Research Council. GHCN and GRDC data are freely available from http://dx.doi.org/10.7289/V5D21VHZ and https://www.bafg.de/GRDC/, respectively.
References (68)
- et al.
A global-scale investigation of trends in annual maximum streamflow
J. Hydrol.
(2017) - et al.
Evidence for global runoff increase related to climate warming
Adv. Water Resour.
(2004) - et al.
An input–output analysis of Australian water usage
Water Policy
(2001) - et al.
Climatic and anthropogenic factors affecting river discharge to the global ocean, 1951–2000
Glob. Planet. Change
(2008) - et al.
Evaluation of outputs from automated baseflow separation methods against simulated baseflow from a physically based, surface water-groundwater flow model
J. Hydrol.
(2012) - et al.
Long range streamflow and world continental runoff fluctuations since the beginning of this century
J. Hydrol.
(1987) - et al.
Characterizing the responses of local floods to changing climate in three different hydroclimatic regions across the United States
Adv. Water Resour.
(2021) - et al.
Influence of changes in rainfall and soil moisture on trends in flooding
J. Hydrol.
(2019) - et al.
Evidence of shorter more extreme rainfalls and increased flood variability under climate change
J. Hydrol.
(2021) - et al.
The global precipitation climatology project (GPCP) monthly analysis (new version 2.3) and a review of 2017 global precipitation
Atmosphere
(2018)