Climate model data are increasingly used to drive hydrological models, to assess the possible impacts of climate change on river flows. Hydrological models often require potential evaporation (PE) from vegetation, alongside precipitation, but PE is not usually output by climate models so has to be estimated from other meteorological variables. Here, the Penman–Monteith formula is applied to estimate PE using data from a 12 km Regional Climate Model (RCM) and a nested very high resolution (1.5 km) RCM covering southern Britain. PE estimates from RCM runs driven by reanalysis boundary conditions are compared to observation-based PE data, to assess performance. The comparison shows that both the 1.5 and 12 km RCMs reproduce observation-based PE well, on daily and monthly time-steps, and enables choices to be made about application of the formula using the available data. Data from Current and Future RCM runs driven by boundary conditions from a Global Climate Model are then used to investigate potential future changes in PE, and how certain factors affect those changes. In particular, the importance of including changes in canopy resistance is demonstrated. PE projections are also shown to vary to some extent according to how aerosols are modelled in the RCMs.

INTRODUCTION

There is increasing concern about the potential impacts of climate change on the hydrological cycle (Stocker et al. 2013). Modelling the possible hydrological impacts is particularly important as changes in the water cycle can affect people and ecosystems both directly (e.g. via changes in water availability and flood frequency) and indirectly (e.g. via changes in food and energy production) (Jiménez Cisneros et al. 2014).

Using Regional Climate Model (RCM) data as an input to hydrological models allows investigation of how climate change may affect river flows (e.g. Ott et al. 2013; Kay & Jones 2012). Many hydrological models require inputs of potential evaporation (PE) from vegetated surfaces, alongside precipitation (Bartholomeus et al. 2015). Unlike precipitation, PE is usually not available directly from RCMs, so has to be estimated from other variables. The meteorological variables that influence PE are temperature, radiation, humidity and wind speed. In addition, vegetation factors such as leaf area and roughness affect the transpiration component (Kay et al. 2013). Many formulae exist for estimating PE, ranging from the physically based Penman–Monteith formula (Monteith 1965) to much simpler empirical formulae like that of Oudin et al. (2005), and the choice can affect the results of subsequent hydrological modelling (e.g. Kay & Davies 2008; Seiller & Anctil 2014). There is much disagreement on the best approach when deriving PE from climate model data for future periods, with concerns about empirical formulae not explicitly including changes in all the influencing variables, but also concerns about data quality when using more complex formulae (see discussion by Kay et al. (2013)).

As part of a recent Natural Environment Research Council (NERC) Changing Water Cycle project, CONVEX, the Met Office ran a very high resolution (1.5 km) RCM for southern Britain, nested in a 12 km RCM driven by global atmospheric reanalysis (ERA-Interim) boundary conditions (1989–2008). They also ran Current (1996–2009) and Future (∼2100 s) climate simulations, with both aerosol climatology and full aerosol modelling setups, nesting the RCMs in a Global Climate Model (GCM). Kendon et al. (2012) found that rainfall in the 1.5 km RCM is more realistic than in the 12 km RCM. In the 12 km RCM, there is a tendency for heavy rain events to be too persistent and widespread, and not heavy enough. Conversely, the 1.5 km RCM has a tendency for heavy rain to be too intense, but it still gives a much better representation of duration and spatial extent.

This paper uses the RCM data from the CONVEX project to estimate PE for short grass, using the Penman–Monteith formula. The following questions are considered: how do RCM estimates of PE compare with observation-based PE; how do the 12 and 1.5 km estimates of PE compare; and how might PE change in the future due to climate change, and what factors influence this PE change?

PE estimates from the ERA-driven RCM runs are compared against observation-based PE from the Met Office Rainfall and Evaporation Calculation System (MORECS) (Hough et al. 1996), as MORECS is the closest to an observational estimate of PE and is widely used by the hydrological community in Britain. Also, the RCM PE estimates are required for an investigation of the use of very high resolution data for hydrological modelling, in which the hydrological models to be used are tuned using MORECS PE along with observed precipitation data (Bell et al. 2012; Crooks et al. 2014). However, MORECS PE is only produced on a 40 × 40 km grid of squares across the UK. The RCM data thus provides the opportunity to estimate PE using a much finer resolution. The comparison includes an assessment of several choices available within the PE estimation method.

Future changes in PE are investigated using the GCM-driven Current and Future RCM simulations. Relatively few studies have looked at potential future changes in PE in Britain, and even fewer have looked at historical changes, either in Britain or globally, but the studies that do exist generally suggest increases (Kay et al. 2013). However, most of these studies have calculated PE changes only from changes in (some of) the meteorological variables; PE can also be affected by increases in the atmospheric concentration of CO2 via changes in stomatal resistance (Bell et al. 2011; Pan et al. 2015). The effect of changes in stomatal resistance is considered here, as is the influence of the method of including aerosols in the RCMs.

Although the focus of this paper is use of RCM data to produce PE estimates that will subsequently be used to drive hydrological models, the issues highlighted will be of wider interest. For example, PE can be an important component in crop modelling (Lovelli et al. 2010) and ecological modelling (Fisher et al. 2011).

METHODOLOGY

The RCM

The 1.5 km RCM is a climate version of the UK Met Office 1.5 km weather forecast model (UKV) (Kendon et al. 2012) but with a smaller domain, spanning southern England and Wales (Figure 1). The 1.5 km RCM lateral boundary conditions are supplied by the 12 km RCM, which is a limited-area (European domain) atmosphere-only version of the Met Office Hadley Centre Global Environmental Model. For the ERA-driven simulations, the 12 km RCM is driven at its lateral boundaries by the latest European Centre for Medium-Range Weather Forecasts Re-Analysis (ERA-Interim) (Dee et al. 2011) for the period 1989 to 2008. For the GCM-driven simulations the 12 km RCM uses boundary conditions from the HadGEM3 GCM, for Current and Future periods. Table 1 summarises the various RCM simulations available, including two versions of the GCM-driven runs with different aerosol formulations (see below). Kendon et al. (2012) found that it takes a few months for soil moisture to spin up in the 1.5 km RCM, therefore only data from January of the first full year of each run are included.

Table 1

Summary of 12 and 1.5 km RCM runs, with run IDs and time periods

  Run ID
 
    
RCM run 12 km 1.5 km Time period Other details 
ERA-driven baseline ajtyr akigd April 1989 – November 2008  
GCM-driven Current (clim) alqtj alxmc May 1996 – November 2009 aerosol climatology 
GCM-driven Future (clim) alqtk alxme as above but for ∼2100s aerosol climatology, RCP 8.5 emissions 
GCM-driven Current (mod) alqtl alxmk May 1996 – November 2009 (12 km). June 1996 – March 2002 (1.5 km) aerosol modelling 
GCM-driven Future (mod) alqtm alxml as above but for ∼2100 s aerosol modelling, RCP 8.5 emissions 
  Run ID
 
    
RCM run 12 km 1.5 km Time period Other details 
ERA-driven baseline ajtyr akigd April 1989 – November 2008  
GCM-driven Current (clim) alqtj alxmc May 1996 – November 2009 aerosol climatology 
GCM-driven Future (clim) alqtk alxme as above but for ∼2100s aerosol climatology, RCP 8.5 emissions 
GCM-driven Current (mod) alqtl alxmk May 1996 – November 2009 (12 km). June 1996 – March 2002 (1.5 km) aerosol modelling 
GCM-driven Future (mod) alqtm alxml as above but for ∼2100 s aerosol modelling, RCP 8.5 emissions 
Figure 1

Map showing the 1.5 km RCM domain (dotted line) and the locations of three MORECS sites used in the analysis (points).

Figure 1

Map showing the 1.5 km RCM domain (dotted line) and the locations of three MORECS sites used in the analysis (points).

For the Current and Future GCM-driven runs, there are both aerosol climatology and fully coupled aerosol modelling runs. The aerosol modelling runs differ from the aerosol climatology runs in that aerosol concentrations are calculated interactively, with advection and deposition of aerosols described. In the case of the 12 km RCM, the aerosols are fully coupled to the microphysics, so that aerosol concentrations determine the number of cloud condensation nuclei. A comparison is made of PE from both pairs of model runs for the 12 km RCM, as data are available for the full period of ∼13 years. The equivalent 1.5 km RCM aerosol modelling runs are only about 5 years in length (Table 1) and are therefore not included in this analysis.

Estimating PE from atmospheric data

PE is generally considered as the amount of water that would be lost to the atmosphere if there were no limits to soil-moisture supply (Kay et al. 2013). The Penman–Monteith method is recommended by the United Nations Food and Agriculture Organization (FAO) for deriving grass reference PE (Pereira et al. 1999) and is used by the UK Climate Projection 09 weather generator (Jones et al. 2009). Penman–Monteith PE (mm/s) is given by 
formula
1
where λ is the latent heat of vaporisation (Jkg−1), Δ is the rate of change of saturated vapour pressure with temperature (kPa °C−1), Rn is the net radiation (Jm−2s−1), ρa is the near surface air density (kgm−3), ca is the specific heat of air (Jkg−1 °C−1), es is the saturation vapour pressure at screen temperature (kPa), ed is the screen vapour pressure (kPa), γ is the psychometric constant (kPa °C−1), ra is the aerodynamic resistance to vapour transfer in the atmosphere (sm−1) and rs is the bulk surface (canopy or bare soil) resistance (sm−1). The saturation vapour pressure es at temperature T (°C) is given by 
formula
2
so Δ= des/dT =17.27 × 237.3 es(T)/(T +237.3)2.
Two different formulations are considered for calculating vapour pressure ed from relative humidity RH (%) and temperature (°C); one uses mean temperature T 
formula
3
and the other uses the min and max temperature (Tmin, Tmax) (Allen et al. 1994) 
formula
4
The aerodynamic resistance ra is calculated from the 10 m wind speed U10 (ms−1) using 
formula
5
which includes a logarithmic correction for wind height (Hough et al. 1996), and surface resistance rs is calculated using 
formula
6
where A =0.7L, L is leaf area index (LAI), rsc is crop resistance and rss is bare soil resistance (100 sm−1) (Hough et al. 1996). MORECS monthly values of rsc and LAI for short grass are used (Hough et al. 1996) (Table 2).
Table 2

Monthly leaf area index (LAI) and crop resistance for Current (rsc_M) and Future (rsc_F) periods, for short grass. LAI and rsc_M are from MORECS, and rsc_F is calculated from rsc_M using Equation (7)

Month LAI Current crop resistance rsc_M Future crop resistance rsc_F 
January 2.0 80 168 
February 2.0 80 168 
March 3.0 60 126 
April 4.0 50 105 
May 5.0 40 84 
June 5.0 60 126 
July 5.0 60 126 
August 5.0 70 147 
September 4.0 70 147 
October 3.0 70 147 
November 2.5 80 168 
December 2.0 80 168 
Month LAI Current crop resistance rsc_M Future crop resistance rsc_F 
January 2.0 80 168 
February 2.0 80 168 
March 3.0 60 126 
April 4.0 50 105 
May 5.0 40 84 
June 5.0 60 126 
July 5.0 60 126 
August 5.0 70 147 
September 4.0 70 147 
October 3.0 70 147 
November 2.5 80 168 
December 2.0 80 168 

The climate model variables used for the calculation of PE are thus 1.5 m temperature, 1.5 m relative humidity, 10 m wind speed and net surface downward longwave and shortwave radiation (which sum to Rn).

Estimating PE for future RCM runs

PE can be affected not just by changes in meteorological inputs, but also by changes in the behaviour of vegetation. In particular, higher CO2 concentrations can lead to plant stomata opening less widely, resulting in higher stomatal resistance, but can also enhance plant growth, leading to a greater leaf area and more stomata (e.g. Bunce 2004; Keenan et al. 2013). Here, following Kruijt et al. (2008), changes in crop resistance are considered but changes in leaf area are not, as it was found by Bunce (2004) that there is a lack of response in LAI due to elevated CO2 concentrations in all functional types, except trees; only grass PE is considered here.

For the Future RCM runs, the equation of Kruijt et al. (2008) is used to estimate appropriate values of grass rsc and then rs. Kruijt et al. (2008) found that the average change in grass and herbal crop conductance (gsc= 1/rsc) per 1 ppm increase in atmospheric CO2 concentration is −9.3 × 10−2% ± 1.5 × 10−2% (under well-watered conditions). The increase in atmospheric CO2 for RCP 8.5 emissions up to year 2100 is 562 ppm, therefore the increase in rsc is 109.5%, or 
formula
7
where rsc_M are the monthly MORECS grass rsc values and rsc_F are the values adjusted for the future climate (Table 2). Surface resistance rs is then calculated from rsc_F using Equation (6). The subsequent rs values compare well with other approaches to estimating future rs, such as taking the average change in rs from members of the UKCP09 RCM ensemble (Murphy et al. 2009) and applying pattern scaling (Mitchell 2003) to allow for differences in emissions scenarios (not shown), and rs changes used by Moratiel et al. (2011).

RESULTS

ERA-driven RCM runs

To choose a calculation of vapour pressure (Equations (3) and (4)), the 12 and 1.5 km estimates of PE are compared with MORECS PE at different timescales, daily and monthly, for three sites in Britain (Figure 1) for the year 1990. The sites were chosen to give spatial coverage across the UK; a southern site (Lyneham), a Midlands site (Nottingham) and a northern site (Galashiels) (not covered by the 1.5 km domain). Figure 2 shows the PE using Equation (3) for vapour pressure; equivalent figures for PE calculated using Equation (4) are similar (not shown). Both estimates of PE are comparable to MORECS, however Equation (3) was chosen as it compares slightly better to MORECS PE.

Figure 2

Comparison of daily and monthly MORECS PE and ERA-driven RCM PE for 1990, for three MORECS sites.

Figure 2

Comparison of daily and monthly MORECS PE and ERA-driven RCM PE for 1990, for three MORECS sites.

To estimate monthly PE, two different averaging methods were compared (for 1990 data): (a) calculating daily PE from daily meteorological variables and then averaging, and (b) calculating monthly averages of the meteorological variables and then calculating monthly PE from them. The two methods yield almost identical results (not shown) so daily PE are made and then the monthly average is calculated (the monthly plots in Figure 2 use this method). This test was done as, although daily data were available here, it can be the case that only monthly data are available, or that it is impractical to obtain all the required data at a daily rather than monthly time-step. In such situations, use of monthly data to calculate monthly PE is unlikely to cause problems (as also suggested by Allen et al. (1998)).

Figure 3 presents maps of the seasonal mean PE for MORECS and the 12 and 1.5 km ERA-driven RCMs for 1990–2007. It shows that the RCMs compare well with MORECS and with each other. The comparison looks to be best for spring (MAM) and winter (DJF). The 12 km RCM might be slightly overdoing summer (JJA) PE in the south-east compared to MORECS and the 1.5 km RCM.

Figure 3

Comparison of seasonal mean PE (mm/day) from MORECS and the 12 and 1.5 km ERA-driven RCMs, for 1990–2007.

Figure 3

Comparison of seasonal mean PE (mm/day) from MORECS and the 12 and 1.5 km ERA-driven RCMs, for 1990–2007.

Current and Future RCM runs

Figure 4 shows maps of seasonal mean PE from the Current and Future aerosol climatology RCM runs, and the percentage change between them. The Current PE for the 12 km and 1.5 km RCMs are similar to each other in magnitude and spatial pattern, with the highest PE in the summer in the southeast and the lowest PE in winter, as for MORECS and the ERA-driven runs (Figure 3). The Future PE from the 12 and 1.5 km RCMs is also similar, but consistently higher than Current PE; PE increases across the country and throughout the year. The percentage change in PE for the 1.5 km run looks larger in winter compared to the 12 km run, however the other seasons look more similar between the two resolutions.

Figure 4

Seasonal mean PE (mm/day) under climate change, including the effect of changes in stomatal resistance. The averages are for 1997–2008 for the Current RCM and for 12 years in the 2100s for the Future RCM. The percentage change between the Current and Future PE is also shown.

Figure 4

Seasonal mean PE (mm/day) under climate change, including the effect of changes in stomatal resistance. The averages are for 1997–2008 for the Current RCM and for 12 years in the 2100s for the Future RCM. The percentage change between the Current and Future PE is also shown.

Figure 5 shows the seasonal mean percentage change in PE for three different setups. The left hand column is for the aerosol climatology 12 km RCM with fixed MORECS values of crop resistance (rsc), the middle column is for the aerosol climatology 12 km RCM with adjusted values of future rsc, and the right hand column shows the equivalent plot for the aerosol modelling 12 km RCM with adjusted future rsc. This shows that the percentage changes in seasonal mean PE are much more similar between the aerosol climatology and full aerosol modelling runs (middle and right) than between the adjusted future rsc and fixed MORECS rsc runs (middle and left). For the adjusted future rsc runs, the summer and autumn percentage changes in PE are larger in the aerosol climatology run than the aerosol modelling run, although the winter and spring changes are similar. Not accounting for the change in stomatal resistance gives a much larger increase in PE from the Current to the Future (left column of Figure 5).

Figure 5

Seasonal mean percentage change in PE for the 12 km RCM. Left is fixed MORECS rsc for both Current and Future, with aerosol climatology. Middle is MORECS rsc for Current and adjusted rsc for Future, with aerosol climatology. Right is MORECS rsc for Current and adjusted rsc for Future, with full aerosol modelling.

Figure 5

Seasonal mean percentage change in PE for the 12 km RCM. Left is fixed MORECS rsc for both Current and Future, with aerosol climatology. Middle is MORECS rsc for Current and adjusted rsc for Future, with aerosol climatology. Right is MORECS rsc for Current and adjusted rsc for Future, with full aerosol modelling.

To take a closer look at the difference in PE changes between the RCMs, Figure 6 shows how the monthly mean percentage change in PE varies for the three MORECS sites (Figure 1). As well as re-emphasising that the percentage change in PE using the fixed MORECS rsc values is consistently higher than the other two runs, these plots also show that the percentage change in PE for the full aerosol modelling run is lower than that from the aerosol climatology run, for June to September. For the other months the values are more comparable. Fully modelling the aerosols appears to have the effect of lowering the summer PE.

Figure 6

Comparison of percentage change in mean monthly 12 km RCM PE for aerosol climatology and full aerosol modelling runs for three MORECS sites.

Figure 6

Comparison of percentage change in mean monthly 12 km RCM PE for aerosol climatology and full aerosol modelling runs for three MORECS sites.

Figure 7 shows the seasonal mean percentage change in shortwave (solar) and longwave radiation for the aerosol climatology and aerosol modelling runs. It shows a larger percentage change in summer radiation in the aerosol climatology run (left column) compared to the aerosol modelling run, especially for shortwave radiation. The seasonal mean percentage change in temperature, relative humidity and wind are more similar for the aerosol climatology and aerosol modelling runs (not shown), so the differences seen in the estimated PE are likely to be due to the radiation differences.

Figure 7

Seasonal mean percentage change in net downward surface radiation (left: shortwave, right: longwave) for the aerosol climatology and full aerosol modelling runs of the 12 km RCM.

Figure 7

Seasonal mean percentage change in net downward surface radiation (left: shortwave, right: longwave) for the aerosol climatology and full aerosol modelling runs of the 12 km RCM.

DISCUSSION AND CONCLUSIONS

Estimates of PE from high-resolution (12 and 1.5 km) RCM data have been produced, using the physically based Penman–Monteith formula, with the aim of making PE suitable to use as input for hydrological modelling. Using ERA-driven RCM data, and a comparison with the observation-based MORECS PE, choices were made about the method to use. The comparison shows that both the 1.5 and 12 km RCM PE estimates are very similar to each other and comparable to MORECS, spatially and at selected locations, at daily and monthly time-steps.

PE estimates from Current and Future RCM runs driven by GCM boundary conditions have been used to investigate potential future changes in PE, and how certain factors affect those changes. It is found that the seasonal mean PE and its change are in close agreement between the 1.5 and 12 km models; this is perhaps not surprising because the large scale changes in humidity, temperature and circulation are common to both RCMs (inherited from the driving GCM). It is also found that future PE is likely to be larger than at present, with the largest increases in summer, autumn and (to a lesser extent) winter. This PE change is influenced by the changes in temperature, relative humidity, wind and radiation as well as stomatal influences through increased CO2. Not accounting for the stomatal influences can inflate the future PE estimates, consistent with the findings of Bell et al. (2011) for UK annual and seasonal mean PE, and such differences can affect subsequent flow projections (e.g. Bell et al. 2012; Prudhomme et al. 2014).

PE projections are also shown to vary to some extent according to how aerosols are modelled in the RCMs. The PE differences from aerosol differences are in the summer/early autumn and this could potentially be important for hydrological modelling as autumn PE that is too high could delay the rewetting of soils in the lead-up to the main flood season in Britain (Bayliss & Jones 1993). The smaller percentage changes in PE for the aerosol modelling run appear to be related to smaller changes in radiation. The potential hydrological importance of aerosol concentrations, via their influence on radiation and so evaporation, is also shown by Gedney et al. (2014), who identify a link between solar dimming, due to rising atmospheric concentrations of aerosols from 1980, and increases in runoff.

Use of the Penman–Monteith PE formulation enabled specific application of changes to crop resistance alongside future changes in meteorological variables. This is not possible (at least in such a straightforward way) with most of the simpler PE formulae, where empirically derived coefficients replace many of the factors present in the more physically based formulae (see for example the summary of 17 PE variations given by Oudin et al. (2005)). There are already concerns about the application of empirical formulae under changing climates (e.g. Donohue et al. 2010; Bartholomeus et al. 2015), and the need to also consider changes in canopy resistance is an added complication (Kay et al. 2013). The use of fixed crop coefficients to estimate crop PE from reference PE is an additional factor that requires consideration under climate change (Bartholomeus et al. 2015), both for hydrological modelling and crop modelling, as different crops may react differently to the same change in CO2 concentrations (Kruijt et al. 2008). Even for hydrological models that do not specifically require PE inputs, similar considerations are likely to apply in the model's internal calculation of evaporation when applied under changing climatic conditions.

Future work will involve running CLASSIC-GB (Crooks et al. 2014) with the high-resolution RCM data to investigate the effect of model resolution, and future climate changes, on peak river flows in southern Britain. As the 1.5 km full aerosol modelling run is not full length, the aerosol climatology run will be used to get sufficient length to look at flood frequencies.

ACKNOWLEDGEMENTS

The authors would like to thank Elizabeth Kendon and Richard Jones (Met Office Hadley Centre) for providing the data and help with analysis of results. This work was funded by the Natural Hazards science area of the NERC-CEH Water and Pollution Science programme.

REFERENCES

REFERENCES
Allen
R.
Smith
M.
Perrier
A.
Pereira
L.
1994
An update for the definition of reference evapotranspiration
.
ICID Bulletin of the International Commission on Irrigation and Drainage
43
,
1
35
.
Allen
R. G.
Pereira
L. S.
Raes
D.
Smith
M.
1998
Crop evapotranspiration (guidelines for computing crop water requirements)
.
FAO Irrigation and Drainage Paper No. 56
,
Food and Agriculture Organization (FAO) of the United Nations
,
Rome
,
Italy
.
Bartholomeus
R. P.
Stagge
J. H.
Tallaksen
L. M.
Witte
J. P. M.
2015
Sensitivity of potential evaporation estimates to 100 years of climate variability
.
Hydrol. Earth Syst. Sci.
19
,
997
1014
.
Bayliss
A.
Jones
R.
1993
Peaks-over-threshold flood database: Summary statistics and seasonality
.
121. IH Report
.
Dee
D. P.
Uppala
S. M.
Simmons
A. J.
Berrisford
P.
Poli
P.
Kobayashi
S.
Andrae
U.
Balmaseda
M. A.
Balsamo
G.
Bauer
P.
Bechtold
P.
Beljaars
A. C. M.
van de Berg
L.
Bidlot
J.
Bormann
N.
Delsol
C.
Dragani
R.
Fuentes
M.
Geer
A. J.
Haimberger
L.
Healy
S. B.
Hersbach
H.
Hólm
E. V.
Isaksen
L.
Kållberg
P.
Köhler
M.
Matricardi
M.
McNally
A. P.
Monge-Sanz
B. M.
Morcrette
J.-J.
Park
B.-K.
Peubey
C.
de Rosnay
P.
Tavolato
C.
Thépaut
J.-N.
Vitart
F.
2011
The ERA-Interim reanalysis: configuration and performance of the data assimilation system
.
Quarterly Journal of the Royal Meteorological Society
137
,
553
597
.
Fisher
J. B.
Whittaker
R. J.
Malhi
Y.
2011
ET Come home: potential evapotranspiration in geographical ecology
.
Global Ecology and Biogeography
20
,
1
18
.
Gedney
N.
Huntingford
C.
Weedon
G.
Bellouin
N.
Boucher
O.
Cox
P.
2014
Detection of solar dimming and brightening effects on Northern Hemisphere river flow
.
Nature Geoscience
7
,
796
800
.
Hough
M.
Palmer
S.
Weir
A.
Lee
M.
Barrie
I.
1996
The Meteorological Office Rainfall and Evaporation Calculation System: MORECS version 2.0 (1995)
.
An update to Hydrological Memorandum 45
,
Met Office Hadley Centre
,
Exeter
,
UK
.
Jiménez Cisneros
B. E.
Oki
T.
Arnell
N. W.
Benito
G.
Cogley
J. G.
Döll
P.
Jiang
T.
Mwakalila
S. S.
2014
Freshwater resources
. In:
Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change
(
Field
C. B.
et al.  
eds).
Cambridge University Press
,
Cambridge
,
UK
and New York, NY, USA
, pp.
229
269
.
Jones
P.
Kilsby
C.
Harpham
C.
Glenis
V.
Burton
A.
2009
UK Climate Projections Science Report: Projections of Future Daily Climate for the UK from the Weather Generator
.
Kay
A.
Bell
V.
Blyth
E.
Crooks
S.
Davies
H.
Reynard
N.
2013
A hydrological perspective on evaporation: historical trends and future projections in Britain
.
Journal of Water and Climate Change
4
,
193
208
.
Keenan
T. F.
Hollinger
D. Y.
Bohrer
G.
Dragoni
D.
Munger
J. W.
Schmid
H. P.
Richardson
A. D.
2013
Increase in forest water-use efficiency as atmospheric carbon dioxide concentrations rise
.
Nature
499
,
324
327
.
Kendon
E.
Roberts
N.
Senior
C.
Roberts
M.
2012
Realism of rainfall in a very high-resolution regional climate model
.
Journal of Climate
25
,
5791
5806
.
Lovelli
S.
Perniola
M.
Di Tommaso
T.
Ventrella
D.
Moriondo
M.
Amato
M.
2010
Effects of rising atmospheric CO2 on crop evapotranspiration in a Mediterranean area
.
Agricultural Water Management
97
,
1287
1292
.
Monteith
J.
1965
Evaporation and environment
.
Symp. Soc. Exp. Biol.
19
,
205
234
.
Moratiel
R.
Snyder
R.
Durán
J.
Tarquis
A.
2011
Trends in climatic variables and future reference evapotranspiration in Duero Valley (Spain)
.
Natural Hazards and Earth System Science
11
,
1795
1805
.
Murphy
J.
Sexton
D.
Jenkins
G.
Boorman
P.
Booth
B.
Brown
K.
Clark
R.
Collins
M.
Harris
G.
Kendon
L.
2009
UK Climate Projections Science Report: Climate change projections
.
Met Office Hadley Centre
,
Exeter
,
UK
.
Ott
I.
Duethman
D.
Liebert
J.
Berg
P.
Feldmann
H.
Ihringer
J.
Kunstmann
H.
Merz
B.
Schaedler
G.
Wagner
S.
2013
High-resolution climate change impact analysis on medium-sized river catchments in Germany: an ensemble assessment
.
Journal of Hydrometeorology
14
,
1175
1193
.
Pereira
L.
Perrier
A.
Allen
R.
Alves
I.
1999
Evapotranspiration: concepts and future trends
.
Journal of Irrigation and Drainage Engineering
125
,
45
51
.
Prudhomme
C.
Giuntoli
I.
Robinson
E. L.
Clark
D. B.
Arnell
N. W.
Dankers
R.
Fekete
B. M.
Franzen
W.
Gerten
D.
Gosling
N. H.
Hagemann
S.
Hannah
D. M.
Kim
H.
Masaki
Y.
Satoh
Y.
Stacke
T.
Wada
Y.
Wisser
D.
2014
Hydrological droughts in the 21st century, hotspots and uncertainties from a global multimodel ensemble experiment
.
PNAS
111
,
3262
3267
.
Stocker
T.
Qin
D.
Plattner
G.-K.
Tignor
M.
Allen
S.
Boschung
J.
Nauels
A.
Xia
Y.
Bex
V.
Midgley
P. M.
(eds)
2013
IPCC, 2013: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change
.
Cambridge University Press
,
Cambridge
,
UK
,
New York, NY, USA
, p.
1535
.