Drought can offset potential water use efficiency of forest ecosystems from rising atmospheric CO2

JES1950_proof ■ 19 December 2019 ■ 1/13 journal of environmental sciences xxx (xxxx) xxx

Available online at www.sciencedirect.com 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64

ScienceDirect www.elsevier.com/locate/jes

Perspective

Q7

Q6

Drought can offset potential water use efficiency of forest ecosystems from rising atmospheric CO2 Ning Liu 1,2, Jatin Kala 1, Shirong Liu 2, Vanessa Haverd 3, Bernard Dell 1,2, Keith R.J. Smettem 1, Richard J. Harper 1,2,* 1

Environmental and Conservation Sciences, Murdoch University, Murdoch, WA 6150 Australia Chinese Academy of Forestry, Beijing 10091 China 3 CSIRO Oceans & Atmosphere, ACT 2600 Australia 2

article info

abstract

Article history:

Increasing atmospheric CO2 is both leading to climate change and providing a potential

Received 6 January 2019

fertilisation effect on plant growth. However, southern Australia has also experienced a

Received in revised form

significant decline in rainfall over the last 30 years, resulting in increased vegetative water

6 March 2019

stress. To better understand the dynamics and responses of Australian forest ecosystems

Accepted 5 December 2019

to drought and elevated CO2, the magnitude and trend in water use efficiency (WUE) of

Available online xxx

forests, and their response to drought and elevated CO2 from 1982 to 2014 were analysed, using the best available model estimates constrained by observed fluxes from simulations

Keywords:

with fixed and time-varying CO2. The ratio of gross primary productivity (GPP) to evapo-

Water use efficiency

transpiration (ET) (WUEe) was used to identify the ecosystem scale WUE, while the ratio of

Elevated CO2

GPP to transpiration (Tr) (WUEc) was used as a measure of canopy scale WUE. WUE

Drought

increased significantly in northern Australia (p < 0.001) for woody savannas, whereas there

Forest

was a slight decline in the WUE of evergreen broadleaf forests in the southeast and

Australia

southwest of Australia. The lag of WUEc to drought was consistent and relatively short and stable between biomes (
3 months), but notably varied for WUEe, with a long time-lag (mean of 10 months). The dissimilar responses of WUEe and WUEc to climate change for different geographical areas result from the different proportion of Tr in ET. CO2 fertilization and a wetter climate enhanced WUE in northern Australia, whereas drought offset the CO2 fertilization effect in southern Australia. © 2019 The Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences. Published by Elsevier B.V.

Introduction Globally, rising CO2 levels along with warming and drought have led to changes in the Water use efficiency (WUE) of

~ uelas et al., 2011; Prior et al., vegetation (Niu et al., 2011; Pen 2011; Xiao et al., 2013; Huang et al., 2015; Liu et al., 2015). WUE is the ratio of the rate of carbon fixation by photosynthesis to water loss and generally reflects the trade-off relationship between water and carbon for vegetation. Generally,

* Corresponding author. Environmental and Conservation Sciences, Murdoch University, Murdoch, WA 6150 Australia. E-mail addresses: [email protected] (N. Liu), [email protected] (R.J. Harper). https://doi.org/10.1016/j.jes.2019.11.020 1001-0742/© 2019 The Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences. Published by Elsevier B.V. Please cite this article as: Liu, N et al., Drought can offset potential water use efficiency of forest ecosystems from rising atmospheric CO2, Journal of Environmental Sciences, https://doi.org/10.1016/j.jes.2019.11.020

65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129

JES1950_proof ■ 19 December 2019 ■ 2/13

2 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

Q1

journal of environmental sciences xxx (xxxx) xxx

WUE is calculated as the ratio of gross primary productivity (GPP) to ET (WUEe) at the ecosystem scale (Niu et al., 2011), reflecting the ratio of carbon assimilation to water loss. Since ET includes transpiration (Tr) from vegetation and evaporation from the soil and canopy surfaces, the ratio of GPP to Tr (WUEc) is used to reflect the coupling relationship between water and carbon at the canopy scale (Ito and Inatomi, 2012). WUE has been broadly correlated with rising CO2 levels, in both observational field studies (e.g., Battipaglia et al. (2013); De Kauwe et al. (2013); Yang et al. (2016)) and modeling studies (e.g., De Kauwe et al. (2013) and Knauer et al. (2017)). The increase in WUE under warming and increasing CO2 is a result of an increase in photosynthesis as atmospheric CO2 levels rise and a decrease of transpiration due to partial closure of stomata (Keenan et al., 2013). However, the work of ~ uelas et al. (2011) indicates that the increase in WUE from Pen an elevation in global CO2 levels does not necessarily result in increased growth of vegetation because of water stress or nutrient limitations. Compared to the responses of vegetation to warming and rising CO2 levels, the effect of drought on WUE is more variable. The impacts of drought on WUE vary with drought condition, particularly the severity, length and frequency of the drought (Wu et al., 2015; Huang et al., 2016; Yu et al., 2017). Moderate and short droughts have been found to increase WUE by reducing transpiration rate (Liu et al., 2015), while severe and frequent droughts may decrease WUE due to the severe decline in photosynthesis (McDowell et al., 2008). Regional decline in water availability may offset the growth-enhancing effects of increasing CO2 levels (RubioCuadrado et al., 2018) and reduce the extent to which forests mitigate climate warming (Brzostek et al., 2014). Drought is one of the most severe impacts of climate change, affecting global water resources and carbon mitigation (Zhao and Running, 2010). The first decade of the 21st century is recognized as the driest decade on record in Australia (Van Dijk et al., 2013). On the one hand, the growth of Australia's vegetation contributed to almost 60% of the global carbon sink anomaly in 2011 (Poulter et al., 2014), which was mainly driven by semi-arid ecosystems (Haverd et al., 2016). On the other hand, both forest dieback (Evans et al., 2013) and greening (Smettem et al., 2013) have been observed across some of Australia's forests. The effect of changes in climate on vegetation greenness varies across the Australian continent depending on the vegetation type. The growth of some vegetation types, such as shrublands and grasslands, is strongly related to water availability (precipitation and soil moisture), but other vegetation types, such as forests and woody savannas, do not show a clear response to drought. For instance, both normalized difference vegetation index (NDVI) and leaf area index (LAI) of non-forested vegetation are significantly correlated to terrestrial water storage across northwest Australia (McGrath et al., 2012; Liu et al., 2017), but there is no direct annual relationship between LAI and rainfall for deep rooted forests of southwest Australia (Smettem et al., 2013). In addition, it has been reported that forest growth may be slower under a warmer climate (Prior and Bowman, 2014), which would reduce carbon sequestration and retard recovery after catastrophic disturbances (Bowman et al., 2014). Actual

evapotranspiration across the Australian continent has a positive relationship with rainfall and temperature (Raupach et al., 2013), while soil moisture, which has been experiencing a long-term decline since 2000 (Chen et al., 2014), is considered the main limiting factor of ET in inland Australia (Jung et al., 2010). However, it is still not clear how CO2 and drought affect the WUE of Australia's forest ecosystems. In addition, it has been reported that vegetation does not respond to climate change immediately, instead, there is a certain time lag to the change in a vegetation index after drought (Wu et al., 2015). By studying the relationship between monthly NDVI and standardized precipitationevapotranspiration index (SPEI), Vicente-Serrano et al. (2013) found that NDVI in inland non-forested areas of Australia was strongly related to SPEI at monthly scales from 1981 to 2006 and showed a relatively short time-lag (<6 months). Therefore, it is important to determine whether there is a time lag of WUE to drought and if differences in this response exist between different vegetation types. Recent developments in land surface modelling for the Australian continent now provide accurate simulations of water and carbon fluxes at a high spatial and temporal resolution, constrained by observations (Haverd et al., 2013; Trudinger et al., 2016). These simulations have been used to examine anomalies of carbon storage, but the interactive responses of the coupling relationship between water and carbon to climate and elevated CO2 have not been explored for deep-rooted vegetation. Without a comprehensive understanding of the response of forests to climate change and elevated CO2, it is difficult to predict how future carbon sequestration of Australian ecosystems will be affected under drying and warming climate conditions. The aim of this study is to characterize the responses of WUE in Australian forest ecosystems to drought and increasing CO2 between 1982 and 2014 using estimates of GPP, ET and Tr from best available model simulations with fixed and time-varying CO2 scenarios. Our specific objectives are to: (1) identify the drought severity for the major forest types using annual SPEI; (2) quantify the magnitude and the temporal dynamics of WUE of Australian forest ecosystems; and (3) quantify the impacts of drought and CO2 on WUE.

1.

Methodology

1.1.

BIOS-2.1 model and forcing data

Australian GPP, ET and Tr (1982e2014) were derived using BIOS2 (Haverd et al., 2013; Trudinger et al., 2016) constrained by multiple observation types, and forced using remotely sensed vegetation cover. BIOS2 is a fine-spatial-resolution (0.05 ) offline modelling environment, including a modification of the CABLE biogeochemical land surface model (Wang et al., 2011) incorporating the SLI soil model (Haverd and Cuntz, 2010). Tr from canopy, evaporation from soil, and net canopy photosynthesis were calculated by the surface flux module in CABLE (Wang et al., 2011). At leaf scale, the transpiration for big sunlit leaves and big shaded leaves are linearly related to the bulk stomatal conductance and water vapor

Please cite this article as: Liu, N et al., Drought can offset potential water use efficiency of forest ecosystems from rising atmospheric CO2, Journal of Environmental Sciences, https://doi.org/10.1016/j.jes.2019.11.020

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

JES1950_proof ■ 19 December 2019 ■ 3/13 journal of environmental sciences xxx (xxxx) xxx

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

Q2

model fraction deficits at the leaf surface. The parameters for a single leaf are scaled up to canopy level using total canopy LAI. Canopy photosynthesis and Tr are coupled through stomatal conductance which is calculated based on the BalleBerryeLeuning model (Ball et al., 1987; Leuning, 1995). Evaporation from soil is calculated using equation A23 of Wang et al. (2011). BIOS2 is forced using the daily gridded meteorological data from the Bureau of Meteorology Australian Water Availability Project (AWAP) (Jones et al., 2017) and soil maps from the Digital Atlas of Australian Soils (McKenzie et al., 2000). BIOS2 parameters are constrained and predictions are evaluated using multiple observation sets from across the Australian continent, including streamflow from 416 gauged catchments (http://www.bom.gov.au/waterdata/), eddy flux data (CO2 and H2O) from 12 OzFlux sites, litterfall data, and soil, litter and biomass carbon pools (Haverd et al., 2013). Results for the current study are extracted from BIOS-2.1 (Trudinger et al., 2016) which was updated to use the GIMMS3g FAPAR product (Zhu et al., 2013). Two simulations from BIOSv2.1, with fixed pre-industrial (280 ppm) and time-varying actual atmospheric CO2 concentrations, were used in this study for separating the effects of climate change and CO2 fertilization on WUE. The time-varying CO2 simulation reflects the combination of impacts of changes in climate as well as CO2 on WUE, while the fixed CO2 simulation only reflects the effect of climate. The difference between time-varying CO2 and fixed CO2 simulations reflects the effect of CO2 on WUE in the model scenarios. Both WUEe and WUEc were used to analyse WUE at ecosystem and canopy scales to better understand the responses of forest function to drought and elevated CO2. Spatial variation in WUEe is controlled largely by the Tr fraction of ET (Tr/ET) (because of the importance of soil evaporation). In the study area, Evergreen broadleaf forests (EBF) have a higher Tr/ET (0.83 ± 0.03) than Woody savannas (WAS) (0.68 ± 0.06).

1.2.

Vegetation types and climate zones

In this study, broad definitions of Australian forests (>50% crown cover), including EBF and WSA, were used to investigate the impact of drought and CO2 on trees. The MODIS Land Cover Type (0.05 ) product (MOD12 Q1) (Friedl et al., 2010) was used to classify different vegetation types. EBF and WSA, make up 3.5% and 7.3% of the Australian landsurface, respectively, as illustrated in Fig. 1. In addition, areas that have experienced recent land cover change were identified by comparing the annual land cover type from 2001 to 2014 for each pixel using the MODIS land cover product. About 20% of forested areas have experienced land cover change in this period, and these areas were excluded from the analysis. € ppen climate classiSeven climate zones following the Ko fication method (Peel et al., 2017), were used to study the impacts of regional climates on WUE including: Tropical climate (Afmw), warm Mediterranean climate (Csa), temperate Mediterranean climate (Csb), warm oceanic climate/humid subtropical climate (Cfa) and temperate oceanic climate (Cfb), as illustrated in Fig. 1.

3

1.3. Standardized precipitation-evapotranspiration index (SPEI) The standardized precipitation-evapotranspiration index (SPEI) was used in this study to quantify drought in forested areas. The SPEI is a multi-scalar drought index based on both precipitation and temperature (Vicente-Serrano et al., 2014). It combines the sensitivity of the Palmer Drought Severity Index (PDSI) to changes in evaporation demand (caused by fluctuations and trends in temperature and vapor pressure deficit (VPD)) with the simplicity of calculation and the multi-temporal nature of the standardized precipitation index (SPI) (Vicente-Serrano et al., 2014). Since it has a multi-scalar character, SPEI can be used for measuring drought severity according to its intensity and duration, and for identifying the accumulated lagged effects on vegetation (Vicente-Serrano et al., 2010). The SPEI is calculated by means of a climatic water balance given by the difference between precipitation and potential evapotranspiration (PET). The complete calculation procedure for the SPEI can be found in Vicente-Serrano et al. (2010). Precipitation is derived from AWAP and PET is derived from Australian Landscape Water Balance's PET output calculated with the Penman equation (Penman, 1948) (Donohue et al., 2009; Viney et al., 2015). In this study, 1e48 months SPEI, spanning the period from 1970 to 2014, was calculated in R 3.3.1 (package SPEI version 1.6) using the “log-logistic” distribution and “ub-pwm” fit methods. Droughts in a given year were classified into three categories, according to the 12-month SPEIs through the end of December (McKee et al., 1993). These three drought severity categories were as follows: extreme drought (ED, SPEI < 2), severe drought (SD, SPEI 2 (2, 1.5)), moderate drought (MD, SPEI 2 (1.5, 1)). We note that the precipitation used to compute the SPEI is from a gridded data-set (Jones et al., 2017), which is commonly used for model evaluation purposes (e.g., Kala et al., 2015; Andrys et al., 2015, 2016), and the PET calculations would invariably introduce some uncertainty in the SPEI calculation which we cannot quantify here. However, we note that the same data-sets have been used for SPEI calculations by others (e.g., Ruthrof et al., 2018; Matusick et al., 2018).

1.4.

Trend analysis of WUE

Trends of all water and carbon processes were calculated using a linear regression model from the annual data series. SlopeCO2 and SlopeClimateþCO2 are the slope of the linear regression model for variables with fixed CO2 (Climate) and time-varying CO2 simulations (Climate þ CO2), respectively; therefore, the trend of variables due to climate change, SlopeClimate, is calculated as the difference between SlopeClimateþCO2 and SlopeCO2. The contribution of climate (ContributionClimate ) and CO2 (ContributionCO2 ) on each variable is calculated based on the trend of those variables. SlopeCO2 ContributionCO2 ¼ 100 SlopeCO2 þ SlopeClimate ContributionClimate ¼ 100

SlopeClimate SlopeCO2 þ SlopeClimate

(1)

(2)

Please cite this article as: Liu, N et al., Drought can offset potential water use efficiency of forest ecosystems from rising atmospheric CO2, Journal of Environmental Sciences, https://doi.org/10.1016/j.jes.2019.11.020

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

JES1950_proof ■ 19 December 2019 ■ 4/13

4 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

journal of environmental sciences xxx (xxxx) xxx

Fig. 1 e Vegetation types from the MODIS Land Cover Type product (MOD12 Q1). Areas that experienced land cover change from 2001 to 2014 were omitted. (Evergreen broadleaf forests (EBF) and Woody savannas (WSA). Letters and lines on the map are climate zone abbreviations and their boundaries: Tropical climate (Afmw), warm mediterranean climate (Csa), temperate mediterranean climate (Csb), warm oceanic climate/humid subtropical climate (Cfa) and temperate oceanic climate (Cfb) (Peel, et al., 2017).).

1.5.

Assessment of the lag-effect of drought on WUE

Lagged effects of drought in this study were assessed with multiple scales of SPEI and a standardized anomaly index of WUE. The anomaly index of WUE (WUESAI) was standardized by the mean and standard deviation of the monthly WUE from 1982 to 2014 and used to analyse the response of WUE to drought, using the Standardized anomaly index (SAI) (Nicholson, 1983): SAIðm;nÞ ¼

Xðm;nÞ  Xm sdðxm Þ

(3)

where, m is month, n is year, Xðm;nÞ is the X in month m of year n, Xm is the X time series in month m over 33 years, sd is the standard deviation, xm is the average of X in month m over 33 years. To calculate the time-lag effects of drought on WUE, at each pixel, the Pearson's correlation coefficient between monthly WUEe and WUEc and 1 to 48 monthly time scale SPEI was calculated. Therefore, there will be 48 correlation coefficients for each pixel, which reflects to 1e48 months' SPEI. Then the maximum correlation of each pixel was selected and the corresponding month of SPEI is defined as the time-lag of drought on WUE.

2.

Results

2.1.

Droughts in Australia's forests from 1982 to 2014

Australia's forested areas have been affected by severe drought (mean annual SPEI < 1) since 1982, although the frequency and severity of drought has varied considerably. This is illustrated in Fig. 2 showing the number of drought years computed using the mean annual SPEI, which was derived from 12-month scale monthly SPEI to reflect the annual drought condition. Northern Australia experienced more frequent droughts (more than 6 years in the 32 years record) during the study period than the southern area. For different vegetation types, a large proportion of WSA (77%) experienced frequent droughts (more than 6 years in the 32 years record) since 1982, followed by EBF (53%). From 1982 to 2014, there were two expansive drought periods, including 1988e1995 and 2002e2007, when, on average, more than 25% of the area suffered droughts for periods within them (Fig. S1). Since 2000, forests in Australia experienced more severe (SPEI < 1.5) and aerially expansive (>25%) droughts for periods within them, such as droughts in 2002, 2003, 2006, 2007 and 2014. In 2003 more than 60% of Australia's forest was

Please cite this article as: Liu, N et al., Drought can offset potential water use efficiency of forest ecosystems from rising atmospheric CO2, Journal of Environmental Sciences, https://doi.org/10.1016/j.jes.2019.11.020

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

JES1950_proof ■ 19 December 2019 ■ 5/13 journal of environmental sciences xxx (xxxx) xxx

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

5

Fig. 2 e Drought frequency (percentage of total years with SPEI < ¡1) from 1982 to 2014 in Australia. Climate zone abbreviations: Tropical climate (Afmw), warm mediterranean climate (Csa), temperate mediterranean climate (Csb), warm oceanic climate/humid subtropical climate (Cfa) and temperate oceanic climate (Cfb).

affected by drought, with about 20% of the area in severe drought (Fig. S1). Fig. 1 shows the annual accumulated drought area (%) for different forest types and climate zones in Australia from 1982 to 2014. It can be seen that more frequent and expansive droughts were observed in the southwest of Australia (SWAU) including Csa and Csb climate zones and the southeast of Australia (SEAU) including Cfa and Cfb climate zones from 2001 to 2014 as compared to the previous three decades, while the opposite trend was observed in the Afmw climate zone in the north of Australia (NAU). There was a significant (p < 0.05) increase in SPEI in NAU, while a marked (p < 0.05) decline of SPEI was found in Csa, Csb's WSA, and Cfb's EBF (Fig. S2). In comparison with droughts before 2000, more EBF were affected by moderate and severe droughts from 2001 to 2014. In 2003 and 2007, the mean annual SPEI was less than 1 and about half of EBF was affected. Furthermore, the EBF in SWAU experienced expansive drought in 1987, 1995, 2001, 2007, 2010 and 2011, while the most severe droughts in EBF in SEAU were in 2003, 2006, 2007, 2008, 2009 and 2014. The spatial distribution of areas experiencing drought is shown in Fig. S3, and only the three most severe and expansive droughts and the wettest year since 1982 (2011) have been shown for clarity. Prior to 2000, the most extensive drought occurred in 1992 when more than 90% EBF, and 75% of WSA in NAU were affected (Fig. 3). After 2000, the most expansive drought (more than 50% of WSA affected) occurred in 2002

and 2003 for WSA. In 2003, a large area of extreme drought was observed for WSA in the central eastern part of Queensland and SEAU. Additionally, extensive droughts for EBF were observed in 2001 in the SWAU. Since 2005, there has been no extensive drought in NAU. By contrast, in 2007, about 40% of EBF and WSA in the south of Australia experienced moderate to severe droughts. Although 2011 was the wettest year since 1982 (Fig. S3), 75% EBF still experienced moderate to severe droughts in SWAU. Droughts mainly occurred in eastern Australia in 2014. Overall, severe droughts occurred more extensively and frequently in EBF and WSA forests in the south of Australia from 2001 to 2014 compared to the period of 1982e2000. Fig. 4 shows the spatial distribution of mean annual WUE in Australia, using both WUEe and WUEc with a time-varying CO2 simulation. Overall, WUEe was spatially consistent with WUEc. However, WUEc for each vegetation type was much larger than that of WUEe as a result of the different ratios of Tr to ET. The mean annual WUE of WSA was lower than WUE for EBF. In addition, WUE increased from the south to the north of the continent for EBF and WSA. Table 1 provides the averages and standard deviations of WUE and its components for each vegetation type and region from 1982 to 2014. On average, EBF forests had the highest mean annual WUEe (2.97 ± 0.33 g C kg1 H2O) followed by WSA (2.36 ± 0.16 g C kg1 H2O). Although a similar spatial pattern of mean annual WUEc was observed, there was a smaller difference in WUEc

Please cite this article as: Liu, N et al., Drought can offset potential water use efficiency of forest ecosystems from rising atmospheric CO2, Journal of Environmental Sciences, https://doi.org/10.1016/j.jes.2019.11.020

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

JES1950_proof ■ 19 December 2019 ■ 6/13

6 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

journal of environmental sciences xxx (xxxx) xxx

Fig. 3 e The total drought area (percentage of area with SPEI < ¡1) for different forest types and climate zones in Australia from 1982 to 2014. Climate zone abbreviations: Tropical climate (Afmw), warm mediterranean climate (Csa), temperate mediterranean climate (Csb), warm oceanic climate/humid subtropical climate (Cfa) and temperate oceanic climate (Cfb). Forest types: Evergreen broadleaf forests (EBF) and Woody savannas (WSA).

Fig. 4 e Spatial distribution of mean annual ecosystem water use efficiency (WUEe, g C kg¡1 H2O) (a) and canopy water use efficiency (WUEc, g C kg¡1 H2O) (b) over the period 1982e2014 across forested areas of Australia. Climate zone abbreviations: Tropical climate (Afmw), warm mediterranean climate (Csa), temperate mediterranean climate (Csb), warm oceanic climate/ humid subtropical climate (Cfa) and temperate oceanic climate (Cfb).

between vegetation types as compared to WUEe. The highest WUEc was found for WSA (3.58 ± 0.32 g C kg1 H2O) in Cfb, while the lowest was observed for EBF (2.95 ± 0.07 g C kg1 H2O) in Csb. The difference between WUEe and WUEc was a result of the different patterns of the ratios of Tr to ET for different vegetation types. EBF had the highest ratio of Tr to ET (>0.8), followed by WSA (about 0.6) (Table 1).

Fig. 5 shows the linear trends of WUEe, WUEc and their components (GPP, ET and Tr) derived from linear regression from the annual data series from both fixed CO2 (Climate) and time-varying CO2 simulations (Climate þ CO2). Overall, contrasting trends in WUE were observed in the north of Australia and the south of Australia for the time varying CO2 simulation. There was a significant decline of WUE in the south of

Please cite this article as: Liu, N et al., Drought can offset potential water use efficiency of forest ecosystems from rising atmospheric CO2, Journal of Environmental Sciences, https://doi.org/10.1016/j.jes.2019.11.020

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

JES1950_proof ■ 19 December 2019 ■ 7/13

7

journal of environmental sciences xxx (xxxx) xxx

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

Table 1 e Comparisons for each vegetation type of mean annual gross primary productivity (GPP), evapotranspiration (ET), transpiration (Tr), ecosystem water use efficiency (WUEe) and canopy water use efficiency (WUEc) over the period 1982e2014. Vegetation type EBF

Region

Climate zone

ET (mm d1)

Tr (mm d1)

GPP (g C m2 d1)

NAU SEAU

Afmw Cfa Cfb Csa Csb Afmw Cfa Cfb Csa Csb

2.77 ± 0.19 2.3 ± 0.28 1.94 ± 0.32 2.31 ± 0.09 2.05 ± 0.17 2.24 ± 0.43 1.79 ± 0.24 1.73 ± 0.24 1.93 ± 0.23 1.42 ± 0.39

2.25 ± 0.34 1.97 ± 0.33 1.65 ± 0.36 1.95 ± 0.1 1.61 ± 0.34 1.29 ± 0.39 1.16 ± 0.29 1.14 ± 0.27 1.46 ± 0.24 0.86 ± 0.49

6.87 ± 0.97 6.32 ± 0.81 5.71 ± 0.87 5.73 ± 0.22 5.28 ± 0.91 3.8 ± 1.15 3.72 ± 0.95 3.89 ± 0.88 4.52 ± 0.56 2.96 ± 1.45

SWAU WSA

NAU SEAU SWAU

WUEe (g C kg1 H2O) 2.48 ± 2.76 ± 2.97 ± 2.48 ± 2.57 ± 1.67 ± 2.05 ± 2.24 ± 2.36 ± 1.98 ±

0.33 0.3 0.33 0.08 0.37 0.38 0.38 0.43 0.16 0.55

WUEc (g C kg1 H2O) 3.07 ± 3.25 ± 3.53 ± 2.95 ± 3.32 ± 2.97 ± 3.22 ± 3.45 ± 3.15 ± 3.58 ±

0.22 0.23 0.32 0.07 0.22 0.29 0.21 0.26 0.17 0.32

SEAU is southeast of Australia, SWAU is southwest of Australia and NAU is north of Australia. (Evergreen broadleaf forests (EBF) and Woody savannas (WSA). Letters and lines on the map are climate zone abbreviations and their boundaries: Tropical climate (Afmw), warm mediterranean climate (Csa), temperate mediterranean climate (Csb), warm oceanic climate/humid subtropical climate (Cfa) and temperate oceanic climate (Cfb); Regions: southwest of Australia (SWAU), north of australia (NAU), and southeast of Australia (SEAU).

Australia but a substantial increase in the north of Australia (Fig. S4). For the time varying CO2 simulation (Climate þ CO2 scenario in Fig. 6), there was an increasing trend of WUE in northern Australia from 1982 to 2014 but a decreasing trend in southern Australia (Fig. S2). Although the trend in annual

WUEc across the entire EBF area was not statistically significant (p > 0.05), the annual WUEe showed a significant increasing trend in Afmw and Csa, with increasing rates of 0.004 g C kg1 H2O yr2 (p < 0.05) and 0.0036 g C kg1 H2O yr2 (p < 0.001), respectively (Fig. 5). Besides, a significant decline of GPP in EBF was observed both in Cfa and Cfb areas, while a

Fig. 5 e Slope derived from the linear regression of ecosystem water use efficiency (WUEe), canopy water use efficiency (WUEc) and their components with fixed CO2 (Climate) and time-varying CO2 (Climate þ CO2) simulations from 1982 to 2014. The *, **, and *** mean a significant trend p < 0.05, p < 0.01 and p < 0.001, respectively. Climate zone abbreviations: Tropical climate (Afmw), warm mediterranean climate (Csa), temperate mediterranean climate (Csb), warm oceanic climate/humid subtropical climate (Cfa) and temperate oceanic climate (Cfb). Forest types: Evergreen broadleaf forests (EBF) and Woody savannas (WSA). Please cite this article as: Liu, N et al., Drought can offset potential water use efficiency of forest ecosystems from rising atmospheric CO2, Journal of Environmental Sciences, https://doi.org/10.1016/j.jes.2019.11.020

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

JES1950_proof ■ 19 December 2019 ■ 8/13

8 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

journal of environmental sciences xxx (xxxx) xxx

Fig. 6 e Contribution (%) of climate and increasing CO2 on ecosystem water use efficiency (WUEe) and canopy water use efficiency (WUEc). The label is the dominant factor; only significant (p < 0.05) factors are colored in this figure. The same direction means that the contribution of the CO2 and climate on the variable is both negative or both positive. Climate zone abbreviations: Tropical climate (Afmw), warm mediterranean climate (Csa), temperate mediterranean climate (Csb), warm oceanic climate/humid subtropical climate (Cfa) and temperate oceanic climate (Cfb). Forest types: Evergreen broadleaf forests (EBF) and Woody savannas (WSA); Regions: southwest of Australia (SWAU), north of australia (NAU), and southeast of Australia (SEAU).

substantial decreasing trend in ET for EBF was observed in Afmw and Csa. For WSA, a significant increasing trend in WUE was observed in Afmw, while a significant increase of WSA was found in Cfa for WUEe and in Csa for WUEc. Moreover, significantly increased trends of WUE and carbon sequestration for WSA were observed in Afmw, while significant decline trend of all both water and carbon processes were observed in Csa. When fixed CO2 was used (Climate scenario in Fig. 5), a significant declining trend in both WUE and GPP was found in most of the EBF. Moreover, a more dramatic decline of WUE in EBF was observed in SEAU as compared to SWAU. In WSA, the response of WUEe differed from that of WUEc. Climate change resulted in a significant decline of WUEc across most of vegetation types, while a significant increasing trend in WUEe was observed in Afmw. The difference between fixed CO2 and time-varying CO2 simulations reflects the different contribution of CO2 and climate in WUE (Fig. 6). In general, increasing CO2 led to an increase of both WUEe and WUEc of WSA across Australia and an increase of EBF in NAU, while drought offset the CO2 fertilization effect on EBF in southwest and southeast Australia (Fig. 6). Climate was the dominant factor for GPP and ET or Tr for most of the regions; however, climate showed a negative contribution to GPP but positive impacts on ET or Tr in Afmw's WSA and Csb's EBF. Conversely, negative impact of CO2 on ET or ET was found in Cfb's EBF.

2.2.

Correlation between drought and WUE

The results in this section report the correlations between drought and WUE for the fixed CO2 and time-varying CO2 simulations. Overall, WUEc over Australia and WUEe of eastern Australia were significantly correlated to SPEI (Table 2). Additionally, WUEc was generally more sensitive to drought than WUEe in most of the Australian ecosystems, because it had a shorter time lag. Table 2 shows the Pearson's correlation coefficient of monthly WUE to SPEI for the main vegetation types and climate zones and the spatial distribution of the time-lag effects of droughts is shown in Fig. S5. WUEc was significantly correlated with SPEI (p < 0.05) in more than 85% of the study areas in all regions. SPEI and WUEc exhibited the most robust relationship in WSA for the north of Australia. Other relatively weak correlations were found for EBF in southeast and southwest Australia. WUEc of most vegetation, except WSA in the southeast of Australia, was positively related to SPEI, with more than 70% showing an accumulated drought time lag less than 3 months (Fig. S5). In contrast, a negative relationship between WUEc and SPEI was observed in WSA at the SEAU. The time-lag effects of droughts on WUEe were far more inconsistent than that of WUEc (Table 2). Only about 60% of Australian forested areas had a significant correlation with drought. For each forest type, the highest proportion of EBF, approximately 79%, showed substantial lag effects of drought, followed by WSA with a proportion of 70%. Overall, forested

Please cite this article as: Liu, N et al., Drought can offset potential water use efficiency of forest ecosystems from rising atmospheric CO2, Journal of Environmental Sciences, https://doi.org/10.1016/j.jes.2019.11.020

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

JES1950_proof ■ 19 December 2019 ■ 9/13

9

journal of environmental sciences xxx (xxxx) xxx

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

Table 2 e Comparisons for each vegetation type of the Pearson's correlation coefficient between monthly SPEI and ecosystem water use efficiency (WUEe) and canopy water use efficiency (WUEc) over the period 1982e2014. Vegetation type

EBF

Region

Climate zone

NAU SEAU

Afmw Cfa Cfb Csa Csb Afmw Cfa Cfb Csa Csb

SWAU WSA

NAU SEAU SWAU

a

Scale (months)

WUEe

WUEc

WUEe

WUEc

0.37a 0.48a 0.37a 0.17 0.12 0.26a 0.33a 0.38a 0.28a 0.27a

0.61a 0.51a 0.51a 0.52a 0.52a 0.62a 0.21a 0.29a 0.37a 0.44a

5 3 3 12 9 1 8 5 1 1

1 1 1 1 1 3 11 12 1 1

Means p < 0.05 in Pearson's method. Climate zone abbreviations: Tropical climate (Afmw), warm Mediterranean climate (Csa), temperate Mediterranean climate (Csb), warm oceanic climate/humid subtropical climate (Cfa) and temperate oceanic climate (Cfb). Forest types: Evergreen broadleaf forests (EBF) and Woody savannas (WSA); Regions: southwest of Australia (SWAU), north of Australia (NAU), and southeast of Australia (SEAU)..

areas in Australia showed an extended length of drought time lag to WUEe, with an average time lag of 10 months. Most of EBF in SWAU were not significantly correlated to drought. A short time lag (3 months) of WUEe was observed in parts of EBF in southeastern Australia. Similar to WUEc, long lag-time effects (>5 months) were observed in parts of WSA in southeast Australia. However, WUEe was positively related to SPEI in WSA in southeastern Australia. In contrast, WUEe of WSA in SWAU was significantly correlated to drought with a 1 month accumulated time lag.

3.

Correlation coefficient

Discussion

This study has found that increasing CO2 has a positive effect on WUE in Australian ecosystems, a result previously reported by Huang et al. (2015). Here the result was found by comparing fixed CO2 and time-varying CO2 simulations, with similar positive effects of elevated CO2 found across different vegetation types. Elevated CO2 levels directly increase photosynthesis and indirectly decrease ET and Tr by partial closure of stomata (Fig. 6) (Keenan et al., 2013). The interactive effects of precipitation changes and elevated CO2 resulted in contrasting trends of WUE between northern and southern Australia.

3.1. Fertilization effects of elevated CO2 and wetter climate increases WUE in the north WSA in Afmw generally had a strong positive relationship between carbon and water processes and water supply (Table 2), showing an increasing trend in all variables (Fig. 5). Although elevated CO2 decreases both ET and Tr of WSA by reducing the stomatal conductance, Tr showed a significant increase due to the wetter climate. The substantial rise in WUE of WSA in Afmw suggests that GPP of woodland responds more rapidly to climate change and elevated CO2 than ET or Tr (Fig. 5 and 6). Moreover, opposite contributions of climate on ET and Tr are shown in Fig. 6, which suggests that the wetter climate resulted in the decline of soil evaporation in the Afmw. This decline of soil evaporation might be because humidity

decreases incoming shortwave radiation. The opposite responses of Tr than ET to climate change resulted in the steeper increase of WUEe than WUEc in WSA (Fig. 5).

3.2. Drought magnitude determined the trend of WUE in the south The direction of WUE varied significantly with vegetation types as a result of the different responses of water and carbon processes to drought and elevated CO2. The severe drying trend increased WUE of WSA in Csa, but decreased the WUE of EBF in Cfb. The significant decline of ET and Tr due to drought and elevated CO2 led to a slight increase of WUE in WSA. WSA is mainly distributed in Cfa and Csa regions, and Csa experienced a more severe drying trend than Cfa (Fig. S2). Even though drought significantly reduced GPP of WSA in Csa, the combined effects of elevated CO2 and drought led to more dramatic relative decreases of ET and Tr than GPP. Other than the drought, the south of Australia also experienced a significant warming trend, which further contributed to the decline of ET or Tr by decreasing stomatal conductance (Lemordant et al., 2018). Meanwhile, De Kauwe et al. (2019) found that extreme heat dramatically decreased GPP in two eddy flux sites in Australia. GPP of EBF dominated the response of WUE to drought in Cfb and Csb. In Cfb, the drying climate resulted in a significant decline of GPP, while Tr and ET increased during the study period. The increase in ET is due to the greening of the forest (Smettem et al., 2013). Sawada and Koike (2016) emphasized that although the vegetation biomass may decrease in a drying environment, vegetation traits such as the carbon allocation strategy may help maintain the greening trend and evapotranspiration. Moreover, a more severe decline in GPP is observed in Cfb than in the Csb. This suggested that EBF in the southeast was more sensitive to drought than in the southwest. Similarly, stronger relationships between vegetation indices (NDVI and LAI) and rainfall and soil moisture have been reported for EBF in southeast Australia as compared to the southwest (Liu et al., 2017). The weak relationship

Please cite this article as: Liu, N et al., Drought can offset potential water use efficiency of forest ecosystems from rising atmospheric CO2, Journal of Environmental Sciences, https://doi.org/10.1016/j.jes.2019.11.020

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

JES1950_proof ■ 19 December 2019 ■ 10/13

10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

journal of environmental sciences xxx (xxxx) xxx

between SPEI and WUE in Csb suggests that these ecosystems did not actively respond to changes in drought over long timescales which were also observed in EBF, woody savannas, and open shrubland in the Northern Hemisphere (Huang et al., 2016). EBF in Csb may have access to dry season water sources other than rainfall, such as stores of deep soil water (Harper et al., 2009; Jung et al., 2010) and groundwater supplies (Smettem et al., 2013). The short-accumulated time scale of SPEI (<3 months) also suggests a contribution of soil moisture to WUE (Chen et al., 2014). Similarly, time-lag effects of soil moisture on vegetation have been found at a continental scale in Australia by Chen et al. (2014). They reported that mainland vegetation had a range of temporal scales (0e5 months), while faster responding (<2 months) areas were located in the east and southwest, in areas dominated by subtropical and temperate grasslands and rain-fed agriculture (both pasture and crops). At the global scale, Wu et al. (2015) also found that the time-lag effects of the vegetation responses significantly affected vegetation growth at global scales using NDVI. Anderegg et al. (2015) found that even though extreme drought did not induce an immediate reduction of forest carbon because of the timelag of drought, it may result in the reallocation of carbon in roots and leaves, and thus ultimately affect the carbon sequestration of forests. In contrast, Walden et al. (2019) found that severe drought caused the death of eucalypts in south-western Australia with a significant transfer of total carbon from live to dead pools, where it was predisposed to future loss through decomposition and fire.

3.3.

Impacts of drought on forest ecosystems

In southwest Australia, the most severe drought occurred from June 2010 to July 2011 in EBF forests. During the first half of this drought, a higher decrease in GPP than ET was observed, which resulted in a significant decline in WUE. This slight decline of ET was because there was still enough stored soil water for the forest to maintain transpiration (May to October is the wet season in this region). In this region, the forest grows on regolith that is 30 to >100 m in depth, with significant soil water and groundwater storage (Harper et al., 2019). However, during the second summer period, a more dramatic decline in ET than GPP resulted in a significant increase in WUE. Moreover, a more dramatic decline of Tr and ET was observed in Cfa as compared to Cfb, as a result of the extreme heat (0.6 C or 3.5% higher) (Brouwers et al., 2013). This period of severe drought, combined with heat wave conditions, was reported to result in markedly reduced forest growth (Evans et al., 2013), some forest mortality (Brouwers et al., 2013) and reduced storage of carbon (Walden et al., 2019). Although water and carbon processes of EBF in Csb are not initially tightly coupled with drought, when drought conditions continue the soil water and groundwater may be depleted, particularly in areas of shallow regolith (Harper et al., 2009). This results in vegetation becoming solely reliant on annual increments of rainfall (Kinal and Stoneman, 2012), which may result in carbon decline or forest mortality in some areas (Walden et al., 2019). In eastern Australia, the most severe drought was observed between October 2002 and August 2003. We found that this

extensive severe drought resulted in a substantial decline in both water and carbon fluxes. However, the decline of ET and Tr is much faster than GPP, which led to a significant decrease in WUE. These different responses of water and carbon fluxes to drought indicated the resilience and water use strategy of forest under drought. For instance, at two eddy flux sites in northern Queensland, Leuning et al. (2005) found that trees could extract groundwater to maintain high annual ET demand during the drought in 2002e2003. Similarly, Keith et al. (2012) found that the severe drought in 2003 resulted in a significant imbalance in the carbon budget with a 26% reduction in GPP in EBF in the southeast of Australia. Disturbances such as fire, cropping and grazing can dramatically change the original land cover type and ecosystem functions that lead to non-linear relationships between vegetation and climate change (Bradshaw, 2012). Although pixels that experienced land cover change from 2001 to 2014, due to severe burning, cropping or other disturbances, have been excluded from this study, there are still some areas that were affected by frequent fires over the period of the record. Fire was not analysed in this study, but it can be an important factor affecting water and carbon dynamics in a range of ecosystems. Other than water shortage, fire is another primary factor affecting the carbon sequestration of Australian ecosystems (Beringer et al., 2007; Brouwers and Coops, 2016). Fire management such as prescribed fires could also change seasonal carbon sequestration (Scheiter et al., 2015). Similarly, Beringer et al. (2015) reported that varied precipitation and rising CO2, along with fire, are likely to alter the structure and function of savannas in northern Australia in the future. We found that about 30% of forest has been affected at least once by fire since 2002, based on 2.5 km resolution fire frequency derived by the ANU Centre for Water and Landscape Dynamics (Van Dijk et al. 2016). Therefore, the impacts of fire in combination with drought on WUE should be evaluated in the future.

4.

Conclusions

The spatial and temporal patterns of WUE and their responses to drought in Australian forest ecosystems from 1982 to 2014 were studied using the SPEI drought index and two indicators of water use efficiency (WUEe and WUEc). We identified that forest ecosystems have experienced more frequent severe droughts since 2002 in the southwest and southeast Australia than the period from 1982 to 2000, while opposite conditions were observed in northern Australia. Overall, a significant increasing trend in WUE for woody savannas, was found during this period in northern Australia, while WUE declined in the south. This contrasting trend of WUE was a result of the frequent droughts in southern Australia. Droughts had significant impacts on WUE of forest ecosystems in the south of Australia; and GPP and Tr were more sensitive to drought than ET in most forest ecosystems. The spatial difference in the ratio of Tr to ET resulted in the variation of WUEc across different vegetation classes, especially in north Australia. Meanwhile, more than half of the forest area had relatively short and significant accumulative lag effects to drought. Our findings suggested that the wet trend and CO2 fertilization

Please cite this article as: Liu, N et al., Drought can offset potential water use efficiency of forest ecosystems from rising atmospheric CO2, Journal of Environmental Sciences, https://doi.org/10.1016/j.jes.2019.11.020

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

JES1950_proof ■ 19 December 2019 ■ 11/13 journal of environmental sciences xxx (xxxx) xxx

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

resulted in the increase of WUE for WAS in north of Australia, while the negative impact of drought in southern Australia offset the CO2 fertilization effect on WUE for EBF.

Q3

Data accessibility statement All source of the data used in the study has been mentioned in the Material and Methods section. The water and carbon simulations of CABLE mode can be required at CABLE's website https://trac.nci.org.au/trac/cable/wiki/CableRegistration.

Uncited reference Tan et al., 2015, Taylor et al., 2014, Vinet and Zhedanov, 2011.

Declaration of Competing Interest Q4

None declared.

Acknowledgments

Q5

Ning Liu is supported by a Murdoch University PhD Strategy Scholarship and National Key Research and Development Program of China (2018YFC0507305-2). The data process in this study was supported by the Pawsey Supercomputing Centre's cloud service (Nimbus). Jatin Kala is supported by an Australian Research Council Discovery Early Career Researcher Award (DE DE170100102). VH acknowledges support from the Earth Systems and Climate Change Hub, funded by the Australian Government's National Environmental Science Program.

Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.jes.2019.11.020.

references

Anderegg, W.R.L., Schwalm, C., Biondi, F., Camarero, J.J., Koch, G., Litvak, M., et al., 2015. Pervasive drought legacies in forest ecosystems and their implications for carbon cycle models. Science 349, 528e532. https://doi.org/10.1126/science.aab1833. Andrys, J., Lyons, T.J., Kala, J., 2015. Multi-decadal Evaluation of WRF Downscaling capabilities over Western Australia in simulating rain- fall and temperature extremes. J. Appl. Meteor. Climatol. 54, 370e394. https://doi.org/10.1175/JAMCD-14-0212.1. Andrys, J., Lyons, T.J., Kala, J., 2016. Evaluation of a WRF ensemble using GCM boundary conditions to quantify mean and extreme climate for the southwest of Western Australia (19701999). Int. J. Climatol. 36, 4406e4424. https://doi.org/10.1002/ joc.4641. Ball, J.T., Woodrow, I.E., Berry, J.A., 1987. A model predicting stomatal conductance and its contribution to the control of

11

photosynthesis under different environmental conditions. In: Biggins, I. (Ed.), Progress in Photosynthesis Research, IV. Martinus-Nijhoff Publishers, Dordrecht, The Netherlands, pp. 221e224. Battipaglia, G., Saurer, M., Cherubini, P., Calfapietra, C., Mccarthy, H.R., Norby, R.J., et al., 2013. Elevated CO2 increases tree-level intrinsic water use efficiency: Insights from carbon and oxygen isotope analyses in tree rings across three forest FACE sites. New Phytol. 197, 544e554. https://doi.org/10.1111/ nph.12044. Beringer, J., Hutley, L.B., Tapper, N.J., Cernusak, L.A., 2007. Savanna fires and their impact on net ecosystem productivity in North Australia. Glob. Chang. Biol. 13, 990e1004. https:// doi.org/10.1111/j.1365-2486.2007.01334.x. Bradshaw, C.J.A., 2012. Little left to lose: Deforestation and forest degradation in Australia since European colonization. J. Plant Ecol. 5, 109e120. https://doi.org/10.1093/jpe/rtr038. Brouwers, N., Matusick, G., Ruthrof, K., Lyons, T., Hardy, G., 2013. Landscape-scale assessment of tree crown dieback following extreme drought and heat in a Mediterranean eucalypt forest ecosystem. Landsc. Ecol. 28, 69e80. https://doi.org/10.1007/ s10980-012-9815-3. Brouwers, N.C., Coops, N.C., 2016. Decreasing Net Primary Production in forest and shrub vegetation across southwest Australia. Ecol. Indic. 66, 10e19. https://doi.org/10.1016/ j.ecolind.2016.01.010. Brzostek, E.R., Dragoni, D., Schmid, H.P., Rahman, A.F., Sims, D., Wayson, C.A., et al., 2014. Chronic water stress reduces tree growth and the carbon sink of deciduous hardwood forests. Glob. Chang. Biol. 20, 2531e2539. https://doi.org/10.1111/ gcb.12528. Chen, T., de Jeu, R.A.M., Liu, Y.Y., van der Werf, G.R., Dolman, A.J., 2014. Using satellite based soil moisture to quantify the water driven variability in NDVI: A case study over mainland Australia. Remote Sens. Environ. 140, 330e338. https://doi.org/ 10.1016/j.rse.2013.08.022. De Kauwe, M.G., Medlyn, B.E., Zaehle, S., Walker, A.P., Dietze, M.C., Hickler, T., et al., 2013. Forest water use and water use efficiency at elevated CO2 : A model-data intercomparison at two contrasting temperate forest FACE sites. Glob. Chang. Biol. 19, 1759e1779. https://doi.org/10.1111/ gcb.12164. De Kauwe, M.G., Medlyn, B.E., Pitman, A.J., Drake, J.E., Ukkola, A., Griebel, A., et al., 2019. Examining the evidence for decoupling between photosynthesis and transpiration during heat extremes. Biogeosciences 16, 903e916. Donohue, R.J., McVicar, T.R., Roderick, M.L., 2009. Generating Australian potential evaporation data suitable for assessing the dynamics in evaporative demand within a changing climate. In: Water for a Healthy Country National Research Flagship: CSIRO, 203. Evans, B., Stone, C., Barber, P., 2013. Linking a decade of forest decline in the south-west of Western Australia to bioclimatic change. Aust. For. 76, 164e172. https://doi.org/10.1080/ 00049158.2013.844055. Friedl, M.A., Sulla-Menashe, D., Tan, B., Schneider, A., Ramankutty, N., Sibley, A., Huang, X., 2010. MODIS Collection 5 global land cover: Algorithm refinements and characterization of new datasets. Remote Sens. Environ. 114, 168e182. https://doi.org/10.1016/j.rse.2009.08.016. Harper, R.J., Smettem, K.R.J., Carter, J.O., McGrath, J.F., 2009. Drought deaths in Eucalyptus globulus (Labill.) plantations in relation to soils, geomorphology and climate. Plant Soil 324, 199e207. https://doi.org/10.1007/s11104-009-9944-x. Harper, R.J., Smettem, K.R.J., Ruprecht, J.K., Dell, B., Liu, N., 2019. Forest-water interactions in the changing environment of south-western Australia. Ann. For. Sci. 76, 95. https://doi.org/ 10.1007/s13595-019-0880-5.

Please cite this article as: Liu, N et al., Drought can offset potential water use efficiency of forest ecosystems from rising atmospheric CO2, Journal of Environmental Sciences, https://doi.org/10.1016/j.jes.2019.11.020

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

JES1950_proof ■ 19 December 2019 ■ 12/13

12 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

journal of environmental sciences xxx (xxxx) xxx

Haverd, V., Raupach, M.R., Briggs, P.R., Canadell, J.G., Isaac, P., Pickett-Heaps, C., et al., 2013. Multiple observation types reduce uncertainty in Australia's terrestrial carbon and water cycles. Biogeosciences 10, 2011e2040. https://doi.org/10.5194/ bg-10-2011-2013. Haverd, V., Smith, B., Trudinger, C., 2016. Dryland vegetation response to wet episode, not inherent shift in sensitivity to rainfall, behind Australia's role in 2011 global carbon sink anomaly. Glob. Chang. Biol. 22, 2315e2316. https://doi.org/ 10.1111/gcb.13202. Huang, L., He, B., Chen, A., Wang, H., Liu, J., Lu, A., et al., 2016. Drought dominates the interannual variability in global terrestrial net primary production by controlling semi-arid ecosystems. Sci. Rep. 6, 24639. https://doi.org/10.1038/ srep24639. Huang, M., Piao, S., Sun, Y., Ciais, P., Cheng, L., Mao, J., et al., 2015. Change in terrestrial ecosystem water-use efficiency over the last three decades. Glob. Chang. Biol. 21, 2366e2378. https:// doi.org/10.1111/gcb.12873. Jones, D., Wang, W., Fawcett, R., 2017. High-quality spatial climate data-sets for Australia. Aust. Meteorol. Oceanogr. J. 58, 233e248. https://doi.org/10.22499/2.5804.003. Jung, M., Reichstein, M., Ciais, P., Seneviratne, S.I., Sheffield, J., Goulden, M.L., et al., 2010. Recent decline in the global land evapotranspiration trend due to limited moisture supply. Nature 467, 951e954. https://doi.org/10.1038/nature09396. Kala, J., Andrys, J., Lyons, T.J., Foster, I.J., Evans, B.J., 2015. Sensitivity of WRF to driving data and physics options on a seasonal timescale for the southwest of Western Australia. Clim. Dyn. 44, 633e659. https://doi.org/10.1007/s00382-014-2160-2. Keenan, T.F., Hollinger, D.Y., Bohrer, G., Dragoni, D., Munger, J.W., Schmid, H.P., et al., 2013. Increase in forest water-use efficiency as atmospheric carbon dioxide concentrations rise. Nature 499, 324e327. https://doi.org/10.1038/nature12291. Keith, H., van Gorsel, E., Jacobsen, K.L., Cleugh, H.A., 2012. Dynamics of carbon exchange in a Eucalyptus forest in response to interacting disturbance factors. Agric. For. Meteorol. 153, 67e81. https://doi.org/10.1016/ j.agrformet.2011.07.019. Kinal, J., Stoneman, G.L., 2012. Disconnection of groundwater from surface water causes a fundamental change in hydrology in a forested catchment in south-western Australia. J. Hydrol. 472e473, 14e24. https://doi.org/10.1016/j.jhydrol.2012.09.013. Knauer, J., Zaehle, S., Reichstein, M., Medlyn, B.E., Forkel, M., Hagemann, S., et al., 2017. The response of ecosystem wateruse efficiency to rising atmospheric CO2 concentrations: sensitivity and large-scale biogeochemical implications. New Phytol. 213, 1654e1666. https://doi.org/10.1111/nph.14288. Lemordant, L., Gentine, P., Swann, A.S., Cook, B.I., Scheff, J., 2018. Critical impact of vegetation physiology on the continental hydrologic cycle in response to increasing CO2. Proc. Natl Acad. Sci. 115, 4093e4098. Leuning, R., 1995. A critical appraisal of a combined stomatalphotosynthesis model for C3 plants. Plant Cell Environ. 18, 339e355. Leuning, R., Cleugh, H.A., Zegelin, S.J., Hughes, D., 2005. Carbon and water fluxes over a temperate Eucalyptus forest and a tropical wet/dry savanna in Australia: Measurements and comparison with MODIS remote sensing estimates. Agric. For. Meteorol. 129, 151e173. https://doi.org/10.1016/ j.agrformet.2004.12.004. Liu, N., Harper, R.J., Dell, B., Liu, S., Yu, Z., 2017. Vegetation dynamics and rainfall sensitivity for different vegetation types of the Australian continent in the dry period 2002e2010. Ecohydrology 10, e1811. https://doi.org/10.1002/eco.1811. Liu, Y., Xiao, J., Ju, W., Zhou, Y., Wang, S., Wu, X., 2015. Water use efficiency of China's terrestrial ecosystems and responses to drought. Sci. Rep. 5, 13799. https://doi.org/10.1038/srep13799.

Matusick, G., Ruthrof, K.X., Kala, J., Brouwers, N.C., Breshears, D.D., Hardy, G.E.S.J., 2018. Chronic historical drought legacy exacerbates tree mortality and crown dieback during acute heatwave-compounded drought. Environ. Res. Lett. 13 (9), 095002. McDowell, N., Pockman, W.T., Allen, C.D., Breshears, D.D., Cobb, N., Kolb, T., et al., 2008. Mechanisms of plant survival and mortality during drought: Why do some plants survive while others succumb to drought? New Phytol. 178, 719e739. https://doi.org/10.1111/j.1469-8137.2008.02436.x. McGrath, G.S., Sadler, R., Fleming, K., Tregoning, P., Hinz, C., Veneklaas, E.J., 2012. Tropical cyclones and the ecohydrology of Australia's recent continental-scale drought. Geophys. Res. Lett. 39 https://doi.org/10.1029/2011GL050263. McKee, T.B., Doesken, N.J., Kleist, J., 1993. The relationship of drought frequency and duraction to time scales. In: Proceedings, Eight Conference on Applied Climatology, 17e22 January 1993. Anaheim, California, pp. 179e184. Niu, S., Xing, X., Zhang, Z., Xia, J., Zhou, X., Song, B., et al., 2011. Water-use efficiency in response to climate change: From leaf to ecosystem in a temperate steppe. Glob. Chang. Biol. 17, 1073e1082. https://doi.org/10.1111/j.1365-2486.2010.02280.x. Penman, H.L., 1948. Natural evaporation from open water, hare soil and grass. Proc. R. Soc. Lond. A. Math. Phys. Sci. 193, 120e145. https://doi.org/10.1098/rspa.1948.0037. ~ uelas, J., Canadell, J.G., Ogaya, R., 2011. Increased water-use Pen efficiency during the 20th century did not translate into enhanced tree growth. Glob. Ecol. Biogeogr. 20, 597e608. https://doi.org/10.1111/j.1466-8238.2010.00608.x. Poulter, B., Frank, D., Ciais, P., Myneni, R.B., Andela, N., Bi, J., et al., 2014. Contribution of semi-arid ecosystems to interannual variability of the global carbon cycle. Nature 509, 600e603. https://doi.org/10.1038/nature13376. Prior, L.D., Bowman, D.M.J.S., 2014. Big eucalypts grow more slowly in a warm climate: Evidence of an interaction between tree size and temperature. Glob. Chang. Biol. 20, 2793e2799. https://doi.org/10.1111/gcb.12540. Prior, S., Prior, S.A., Runion, G.B., Marble, S.C., Rogers, H.H., Gilliam, C.H., 2011. A Review of Elevated Atmospheric CO2 Effects on Plant Growth and Water Relations: Implications for Horticulture. HortScience 46, 158e162. Raupach, M.R., Haverd, V., Briggs, P.R., 2013. Sensitivities of the australian terrestrial water and carbon balances to climate change and variability. Agric. For. Meteorol. 182e183, 277e291. https://doi.org/10.1016/j.agrformet.2013.06.017. Ruthrof, K.X., Breshears, D.D., Fontaine, J.B., Froend, R.H., Matusick, G., Kala, J., et al., 2018. Subcontinental heat wave triggers terrestrial and marine, multi-taxa responses. Sci. Rep. 8 (1), 13094. Sawada, Y., Koike, T., 2016. Ecosystem resilience to the Millennium drought in southeast Australia (2001e2009). J. Geophys. Res. Biogeosciences 121, 2312e2327. https://doi.org/ 10.1002/2016JG003356. Scheiter, S., Higgins, S.I., Beringer, J., Hutley, L.B., 2015. Climate change and long-term fire management impacts on Australian savannas. New Phytol. 205, 1211e1226. https://doi.org/ 10.1111/nph.13130. Smettem, K.R.J., Waring, R.H., Callow, J.N., Wilson, M., Mu, Q., 2013. Satellite-derived estimates of forest leaf area index in southwest Western Australia are not tightly coupled to interannual variations in rainfall: Implications for groundwater decline in a drying climate. Glob. Chang. Biol. 19, 2401e2412. https://doi.org/10.1111/gcb.12223. Tan, Z.H., Zhang, Y.P., Deng, X.B., Song, Q.H., Liu, W.J., Deng, Y., et al., 2015. Interannual and seasonal variability of water use efficiency in a tropical rainforest: Results from a 9 year eddy flux time series. J. Geophys. Res. 120, 464e479. https://doi.org/ 10.1002/2014JD022535.

Please cite this article as: Liu, N et al., Drought can offset potential water use efficiency of forest ecosystems from rising atmospheric CO2, Journal of Environmental Sciences, https://doi.org/10.1016/j.jes.2019.11.020

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130

JES1950_proof ■ 19 December 2019 ■ 13/13 journal of environmental sciences xxx (xxxx) xxx

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

Taylor, J.E., Ellis, M.V., Rayner, L., 2014. Growth, recruitment and attrition of Eucalyptus tree species in semi-arid temperate woodland. For. Ecol. Manage. 331, 25e34. https://doi.org/ 10.1016/j.foreco.2014.07.023. Trudinger, C.M., Haverd, V., Briggs, P.R., Canadell, J.G., 2016. Interannual variability in Australia's terrestrial carbon cycle constrained by multiple observation types. Biogeosciences 13, 6363e6383. https://doi.org/10.5194/bg-13-6363-2016. Van Dijk, A.I.J.M., Beck, H.E., Crosbie, R.S., De Jeu, R.A.M., Liu, Y.Y., Podger, G.M., et al., 2013. The Millennium Drought in southeast Australia (2001-2009): Natural and human causes and implications for water resources, ecosystems, economy, and society. Water Resour. Res. 49, 1040e1057. https://doi.org/ 10.1002/wrcr.20123. Van Dijk, Albert, I.J.M., Summers, D., 2016. Australia's Environment Explorer. The Australian National University. https://doi.org/10.4225/41/5934faf27397b. http://www.ausenv. online.  pez-Moreno, J.I., 2010. A Vicente-Serrano, S.M., Beguerı´a, S., Lo multiscalar drought index sensitive to global warming: The standardized precipitation evapotranspiration index. J. Clim. 23, 1696e1718. https://doi.org/10.1175/2009JCLI2909.1. Vicente-Serrano, S.M., Lopez-Moreno, J.I., Beguerı´a, S., LorenzoLacruz, J., Sanchez-Lorenzo, A., Garcı´a-Ruiz, J.M., et al., 2014. Evidence of increasing drought severity caused by temperature rise in southern Europe. Environ. Res. Lett. 9, 044001 https://doi.org/10.1088/1748-9326/9/4/044001. Vinet, L., Zhedanov, A., 2011. A “missing” family of classical orthogonal polynomials. J. Phys. A Math. Theor. 44, 660e684. https://doi.org/10.1088/1751-8113/44/8/085201. Viney, N., Vaze, J., Crosbie, R., Wang, B., Dawes, W., Frost, A., 2015. AWRA-L v5.0: technical description of model algorithms and inputs. CSIRO, Australia. https://doi.org/10.4225/08/ 58518bc790ff7.

13

Walden, L.L., Fontaine, J.B., Ruthrof, K.X., Matusick, G., Harper, R.J., Hardy, G.E., 2019. Carbon consequences of drought differ in forests that resprout. Glob. Chang Biol. 25, 1653e1664. https://doi.org/10.1111/gcb.14589. Wang, Y.P., Kowalczyk, E., Leuning, R., Abramowitz, G., Raupach, M.R., Pak, B., et al., 2011. Diagnosing errors in a land surface model (CABLE) in the time and frequency domains. J. Geophys. Res.: Biogeosciences 116 (G1). Wu, D., Zhao, X., Liang, S., Zhou, T., Huang, K., Tang, B., et al., 2015. Time-lag effects of global vegetation responses to climate change. Glob. Chang. Biol. 21, 3520e3531. https:// doi.org/10.1111/gcb.12945. Xiao, J., Sun, G., Chen, J., Chen, H., Chen, S., Dong, G., et al., 2013. Carbon fluxes, evapotranspiration, and water use efficiency of terrestrial ecosystems in China. Agric. For. Meteorol. 182e183, 76e90. https://doi.org/10.1016/j.agrformet.2013.08.007. Yang, Y., Guan, H., Batelaan, O., McVicar, T.R., Long, D., Piao, S., et al., 2016. Contrasting responses of water use efficiency to drought across global terrestrial ecosystems. Sci. Rep. 6, 23284. https://doi.org/10.1038/srep23284. Yu, Z., Wang, J., Liu, S., Rentch, J.S., Sun, P., Lu, C., 2017. Global gross primary productivity and water use efficiency changes under drought stress. Environ. Res. Lett. 12 https://doi.org/ 10.1088/1748-9326/aa5258. Zhao, M., Running, S.W., 2010. Drought-induced reduction in global terrestrial net primary production from 2000 through 2009. Science 329, 940e943. https://doi.org/10.1126/science.1192666. Zhu, Z., Bi, J., Pan, Y., Ganguly, S., Anav, A., Xu, L., et al., 2013. Global data sets of vegetation leaf area index (LAI)3g and fraction of photosynthetically active radiation (FPAR)3g derived from global inventory modeling and mapping studies (GIMMS) normalized difference vegetation index (NDVI3G) for the period 1981 to 2011. Remote Sens. 5, 927e948. https:// doi.org/10.3390/rs5020927.

Please cite this article as: Liu, N et al., Drought can offset potential water use efficiency of forest ecosystems from rising atmospheric CO2, Journal of Environmental Sciences, https://doi.org/10.1016/j.jes.2019.11.020

66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130