Elsevier

Journal of Hydrology

Volume 541, Part B, October 2016, Pages 1042-1056
Journal of Hydrology

Research papers
The analytical derivation of multiple elasticities of runoff to climate change and catchment characteristics alteration

https://doi.org/10.1016/j.jhydrol.2016.08.014Get rights and content

Highlights

  • Developed analytical derivation of multiple elasticities based on Budyko and penman framework.

  • Quantified the contributions of climatic conditions and catchment characteristics to runoff change.

  • Preliminarily assessed the applicability of model extension.

Abstract

The concept of elasticity has been widely employed to quantify the hydrological response to changes in climate and catchments properties. To separate the effect of different climatic variables on runoff, the potential evaporation (E0) elasticity of runoff needs to be presented in term of observed climate variables. To fully reflect the effects of maximum and minimum temperatures and reduce the influence of the correlations of radiation with sunshine duration and relative humidity on the assessment results, we decompose the E0 elasticity into five evaporation-related elasticities (i.e., sunshine duration, maximum and minimum temperature, wind speed and relative humidity) via the first-order differentiation of the FAO 56 Penman equation. As the catchment runoff is frequently affected by the land use/cover change, we also consider changes in catchment characteristics and derive a catchment alteration elasticity based on the Budyko framework. An application was carried out in 30 catchments with widespread climatic types in China. For the two periods (i.e., the baseline period and the changed period) divided by the Pettitt test, the contributions of different climatic variables and land use/cover conditions to runoff change were quantified. In general, the alteration of catchment characteristics and climatic change should be mainly responsible for changes in runoff in water-limited and humid basins, respectively. Although the elasticity of maximum temperature are usually higher than that of minimum temperature, the contributions to runoff change present the opposite direction. Furthermore, additional analysis indicated some overestimation in relative humidity elasticities in the previous studies, further emphasizing the necessity of our extension to alleviate the influence of correlation between climatic variables to the assessment results. Moreover, the results of model performance versus model complexity showed that the choice of model complexity still depends on the goal in the climatic and anthropogenic effects on runoff. If the climatic effect as a whole was of interest, the method based on two parameters elasticities would be more applicable compared to that based on multi-parameter elasticities because the simple model provides higher confidence in modelling the overall change in runoff.

Introduction

Climate change was widely reported in the past decades and is expected to continue (IPCC, 2007). Concurrently, human activities increased due to strong growths in population, including cultivation, irrigation, afforestation, deforestation and urban construction (Brown et al., 2005, Scanlon et al., 2007, Roderick and Farquhar, 2011). Both climate change and increasing human activities intensify the environmental changes that are occurring around the world. Understanding the interactions between climate change, human-induced land use/cover change and the terrestrial water cycle under the changing environment is critical for many research disciplines, such as hydrology, meteorology and ecology (Eagleson, 2002, Rodriguez-Iturbe and Porporato, 2004, Li et al., 2013). Water resources planning and management urgently need to incorporate climate variability and anthropogenic activities to develop sustainable strategies. Therefore, a deceptively simple question - how do climate variability and human activities respectively attribute to the change of annual runoff? - has attracted the attention of hydrologists and water resources managers and planners (e.g., Lane et al., 2005, Siriwardena et al., 2006, Tuteja et al., 2007, St Jacques et al., 2010, Wang et al., 2013a). Traditionally, paired catchment method and physically based method are used to quantitatively estimate the contributions of climate change and human activities to changes in the water cycle. The paired catchment method is considered to be the foundation of studying forestry and vegetation effects on hydrology (Brown et al., 2005). However, paired catchments are not always available due to difficulties in locating suitable controls and are expensive to setup (Fohrer et al., 2005, Wei and Zhang, 2010, Zhao et al., 2010). The physically based approaches (by the means of distributed physically-based hydrological models) are always limited due to the time-consuming model calibration, the requirement of large data sets and high uncertainties in the model structure and parameter estimation (Zhao et al., 2010, Wang et al., 2013a, Xu et al., 2014).

Recently, an approach based on climate elasticity, which was first introduced by Schaake (1990), is being developed actively to assess the climatic and anthropogenic effects on runoff due to its clear physical meanings and simple formulation (e.g., Dooge, 1992, Sankarasubramanian et al., 2001, Fu et al., 2007, Zheng et al., 2009, Yang and Yang, 2011, Roderick and Farquhar, 2011). The climate elasticity of runoff is defined as the proportional change in runoff due to the proportional change in one or more climatic variables (Sankarasubramanian et al., 2001, Fu et al., 2007). During the past two decades, the climate elasticity of runoff has roughly undergone three major expansions. Schaake (1990) defined the precipitation (P) elasticity to runoff (Q) as:εP(P,Q)=dQdPPQand the impacts of P on Q can be evaluated by:dQQ=εp(P,Q)dPP

Based on Eq. (2), a non-parametric approach was introduced by Sankarasubramanian et al. (2001) and found to be robust via Monte Carlo experiments in three basins of the United States when directly identifying the response of runoff to precipitation based on observed long-term hydrometeorological data. Many studies have employed Eq. (2) to investigate the impact of precipitation change on annual runoff (e.g., Sankarasubramanian and Vogel, 2003, Niemann and Eltahir, 2005, Chiew, 2006, Novotny and Stefan, 2007).

Apart from precipitation change, air-temperature change can also have important impacts on runoff, which is particularly true under global warming. To simultaneously assess the impact of precipitation and temperature on runoff, Fu et al. (2007) extended Eq. (2) to consider both precipitation and temperature elasticities by:dQQ=εadPP+εbdTTwhere εa and εb are the P elasticity and air temperature (T) elasticity, respectively, and revealed the critical role of T in regional runoff generation in two typical rivers: the Spokane River Basin in the USA and the Yellow River Basin in China. Fu et al. (2009) further revealed the potentially serious changes in water resources management in the North China Plain under future climate scenarios based on this equation. Obviously, in terms of climate change, integrating the effects of temperature, wind speed, solar radiation and vapor pressure, potential evapotranspiration is more competent than considering only temperature when investigating the influence of climate on hydrological processes, particularly in the case of the widely reported “evaporation paradox” in recent years. Zheng et al. (2009) afterwards proposed a similar expression of climate elasticity of runoff but in terms of precipitation and potential evapotranspiration as:dQQ=εadPP+εbdE0E0where εa and εb are the P elasticity and potential evapotranspiration (E0) elasticity, respectively, and applied it to quantify the elasticity of runoff to climate and catchment characteristics change.

Note that potential evapotranspiration is not directly measured, and temperature is not the only climate variable that affects potential evapotranspiration. Yang and Yang (2011) further evaluated the effects of more climatic variables such as the net radiation and wind speed on runoff by expanding E0 in terms of temperature, net radiation (Rn), wind speed (U) at a height of 2 m above the ground and relative humidity (RH), producing a new equation:dQQ=εadPP+εbdTT+εcdRnRn+εddUU+εedRHRHwhere εa, εb, εc, εd and εe are the elasticities of runoff with respect to P, T, Rn, U and RH respectively. The spatial variations of climate elasticities across China were further characterized in Yang et al. (2014a). Based on this equation, Xu et al. (2014) conducted an attribution analysis to quantify the impacts of climate variation and land use/cover change on the decline of runoff in the Hai River Basin. The major expansions of climate elasticity of runoff and their applications have been summarized in Table 1.

In addition to climate change, the alteration of catchment characteristics can also play a key role in the variability of the flow regime and subsequently the amount of available water resources, particularly in rapidly developing regions, although catchment characteristics were always assumed to remain constant over time in climate-change impact studies. In this study, the authors summary that the Budyko framework with two climatic variables and one catchment-specific parameter, describing the dependence of actual evapotranspiration on energy availability and water availability, made it possible to include the effect of changes in catchment characteristics over time when employing the analytical model to construct elasticities. Roderick and Farquhar (2011) appears to have been the first to present analytical expressions to qualitatively estimate the impact of catchment characteristic alteration and climatic change on runoff change within the Budyko framework. Recently, Liang et al. (2015) quantify the impacts of climate change and ecological restoration on runoff changes in the Loess Plateau of China based on this framework.

With long-standing interest in theoretical aspects of the “Budyko curve”, researchers successively proposed more analytical equations during past decades. Among these equations, two threads of independent theoretical development have attracted many attentions (Roderick and Farquhar, 2011), originating respectively from Russian scholars (Bagrov, 1953, Mezentsev, 1955) in the form ofE=PE0(Pn+E0n)1/n(for the generalized form of Eq. (6), see Choudhury, 1999) and Chinese scholar (Fu, 1981) in the form ofE=P+E0-(Pω+E0ω)1/ω(for the generalized form of Eq. (7), see Zhang et al., 2004), where E is the actual evapotranspiration; and n and ω are the empirical dimensionless parameters, respectively, which play similar roles in their equations to determine the shape of the Budyko curve and reflect the impact of catchment characteristics. Recently, with the theoretically derived Choudhury’s equation (Eq. (6)) based on Fu’s approach, these two equations were reconciled by Yang et al. (2008). Further analysis using the data from numerous catchments in China indicated that the catchment parameters in the two equations are more and less linearly related through ωn+0.72 (Sun, 2007, Yang et al., 2008).

In practice, some issues still need to be addressed in the development of the climatic elasticity, particularly for the impact assessment study. To date, four evaporation-related climatic variables were usually considered in the study of the climatic elasticity of runoff: mean air temperature, radiation, wind speed and relative humidity (see Eq. (5)). However, for the purpose of impact assessment study and with the consideration of data availability, the following issues need to be considered. (1) As the primary representation of climate change, warming trends are actually proved to be more pronounced in minimum temperature (night time temperature) (Alexander et al., 2006, Wang et al., 2013b). Compared to the average air temperature, the extreme temperatures are of more significance to climate change and thus should be considered as key variables when the climate elasticity is used to assess climate change effects on runoff. (2) In FAO Penman-Montieth equation, net radiation (Rn) is computed from sunshine duration (N), relative humidity (RH), maximum temperature (TA) and minimum temperature (TI) and thus correlates with them. Such correlations may cause relative errors when taking Rn, RH, TA and TI as the key variables to estimate the climatic effect on runoff. Furthermore, limited by the cost of measurement equipment and the difficulties of its maintenance and calibration (Hunt et al., 1998, Thornton and Running, 1999, Chen and Li, 2013), the accurate long-term records of solar radiation are not widely available. Therefore, the direct use of radiation in climate-change impact studies might not be widely applicable in practice. (3) In recent studies on the runoff change attribution analysis with elasticity method (e.g., Yang and Yang, 2011, Meng and Mo, 2012), although the effects of evaporation-related climatic variables (e.g., air temperature, wind speed, relative humidity and radiation) on runoff were evaluated and separated, the effect of alteration of catchment characteristics on runoff change was ignored, that is, the interannual changes in catchment characteristics was not considered. Therefore, an improved multiple climate elasticity as well as catchment characteristics elasticity of runoff need to be derived with the consideration of the influence of correlation and its applicability.

From the aspect of model complexity, the performances of different models with different complexity/extension have not been investigated, and the requirements of these models have not been clearly delineated. Although it is not clear how many climatic variables should be theoretically considered in the assessment of the climatic effect on runoff, it is adequate to use the precipitation and evapotranspiration elasticities (see the Eq. (4)) to describe the complex non-linear relationship between the hydrological and climatic regimes based on the normal water balance theory if one only wants to identify the integrated effect of climate change. More importantly, from the mathematical perspective for a given response variable, more assessment errors may be produced if more control variables are used (i.e., more elasticities are considered). This may produce a dilemma regarding additional insights on the effect of climate in term of more climate variables versus possibly more reliable evaluation.

To address these issues, we extend five climatic elasticities into six parameters elasticities by using the widely recommended Penman-Montieth equation in deriving reference evapotranspiration to fully consider the maximum and minimum temperature effect and alleviate its influence on the assessment results from the correlation among climatic variables. That is, the climate elasticity of runoff is considered to be a function of precipitation, sunshine duration, wind speed, relative humidity, minimum and maximum temperature. An analytical framework based on the Budyko hypothesis was subsequently established to decompose the climate-induced and catchment characteristics-induced runoff change. As a case study, the method was applied to 30 hydrological stations in major rivers across China. Moreover, as a preliminary study, we also assessed the applicability and reliability of climate elasticities with different numbers of climate factors by investigating their performances on the quantitative attributions to runoff. This paper is organized as follows. The derivation of multiple elasticity is given in the next section, followed by the results of the analysis and application in 30 catchments. The discussion, including the necessity of multiple elasticities, the discussion on the role of catchment property in runoff change and the uncertainties analysis, are presented before the conclusions.

Section snippets

Elasticity of runoff derived from the Budyko hypothesis equation

The water balance for a catchment can be described by P=E+Q+ΔS. On the annual scale, the transient changes in soil water storage (ΔS) can reasonably be neglected with the assumption that the change in ΔS in a given catchment is from one steady state to another steady state, and can thus be considered equal to zero for a long period. Thus, the normal water balance equation can be simplified as P=E+Q, leading to the differential form:dP=dE+dQ

Eq. (6) can be written as E=f(E0,P,n) to estimate the

Study area and data processing

The elasticity theory and decomposition framework formulated in this study is used to quantitatively assess the effects of the climate and catchment characteristics on runoff in the primary, large river basins across China including the Songhua River Basin, the Liao River Basin, the Hai River Basin, the Yellow River Basin, the Huai River Basin, the Yangtze River Basin, and the Pearl River Basin. The monthly runoff data from 30 hydrological stations in these seven major basins were collected

Long-term trends analysis and breakpoint detection for a hydrometeorological series

Table 3 shows the statistical results of the long-term trends and their significance for hydrometeorological variables based on the nonparametric Mann-Kendall test (Mann, 1945, Kendall, 1975) from 1956 to 2000. Significantly positive trends in TI can also be found in the seven basins (25 catchments for total 30 catchments). However, only a few stations had significantly positive trends in TA. Therefore, the warming processes in the study area were primarily characterized by increasing minimum

The necessity and applicability of multiple elasticities of runoff

Elasticity can well represent the impact of a given change on the normal state of a system (Harman et al., 2011) that is readily calculable and thus has been widely used to assess a hydrological regime with respect to environmental changes in past decades. From 1990, the runoff elasticity has undergone the development process from a single-parameter (e.g., precipitation) climatic elasticity (e.g., Schaake, 1990) to a two-parameter (e.g., precipitation and temperature) climate elasticity (e.g.,

Conclusion

To reflect the effect of maximum and minimum temperature on runoff and avoid the errors from the correlation between radiation, air temperature and relative humidity, a five evaporation-related variables analytical derivation of climate elasticities was proposed in this study to further separate and evaluate the climatic effect on runoff more systematically. Based on this derivation, a decomposition framework with the ability to independently estimate the contributions of climatic and

Acknowledgments

This work was jointly supported by the National Natural Science Foundation of China (51379057), the Fundamental Research Funds for the Central Universities (2015B14114), National “Ten Thousand Program” Youth Talent, the Australian Endeavour Research Fellowship, CSIRO Computational and Simulation Sciences TCP, and QingLan Project. Cordial thanks are extended to the Editor, Professor Geoff Syme, and two anonymous referees for their valuable comments.

References (67)

  • X. Xu et al.

    Attribution analysis based on the Budyko hypothesis for detecting the dominant cause of runoff decline in Haihe basin

    J. Hydrol.

    (2014)
  • H. Yang et al.

    The regional variation in climate elasticity and climate contribution to runoff across China

    J. Hydrol.

    (2014)
  • Y. Yang et al.

    Abrupt change of runoff and its major driving factors in Haihe River Catchment, China

    J. Hydrol.

    (2009)
  • Y. Yokoo et al.

    Investigating the roles of climate seasonality and landscape characteristics on mean annual and monthly water balances

    J. Hydrol.

    (2008)
  • L.V. Alexander et al.

    Global observed changes in daily climate extremes of temperature and precipitation

    J. Geophys. Res.

    (2006)
  • R.G. Allen et al.

    Crop Evapotranspiration: Guidelines for Computing Crop Water Requirements

    (1998)
  • N.A. Bagrov

    Mean long-term evaporation from land surface

    Meteorol. Gidrol.

    (1953)
  • J.L. Chen et al.

    Estimation of monthly average daily solar radiation from measured meteorological data in Yangtze River Basin in China

    Int. J. Climatol.

    (2013)
  • F.H.S. Chiew

    Estimation of rainfall elasticity of streamflow in Australia

    Hydrol. Sci. J.

    (2006)
  • R.J. Donohue et al.

    On the importance of including vegetation dynamics in Budyko’s hydrological model

    Hydrol. Earth Syst. Sci.

    (2007)
  • J.C.I. Dooge

    Sensitivity of runoff to climate change: a Hortonian approach

    Bull. Am. Meteorol. Soc.

    (1992)
  • P.S. Eagleson
    (2002)
  • N. Fohrer et al.

    Assessment of the effects of land use patterns on hydrologic landscape functions: development of sustainable land use concepts for low mountain range areas

    Hydrol. Process.

    (2005)
  • B.P. Fu

    On the calculation of the evaporation from land surface

    Sci. Atmos. Sin.

    (1981)
  • G. Fu et al.

    A two-parameter climate elasticity of streamflow index to assess climate change effects on annual streamflow

    Water Resour. Res.

    (2007)
  • G. Fu et al.

    Decadal climatic variability, trends, and future scenarios for the North China plain

    J. Clim.

    (2009)
  • G. Gutman et al.

    The derivation of the green vegetation fraction from NOAA/AVHRR data for use in numerical weather prediction models

    Int. J. Remote Sens.

    (1998)
  • C.J. Harman et al.

    Functional model of water balance variability at the catchment scale: 2. Elasticity of fast and slow runoff components to precipitation change in the continental United States

    Water Resour. Res.

    (2011)
  • IPCC, 2007. Summary for policymakers. In: Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B.,...
  • M.G. Kendall

    Rank Correlation Methods

    (1975)
  • D. Li et al.

    Vegetation control on water and energy balance within the Budyko framework

    Water Resour. Res.

    (2013)
  • W. Liang et al.

    Quantifying the impacts of climate change and ecological restoration on streamflow changes based on a Budyko hydrological model in China’s Loess Plateau

    Water Resour. Res.

    (2015)
  • H.B. Mann

    Nonparametric tests against trend

    Econometrica

    (1945)
  • Cited by (0)

    View full text