Responses of evapotranspiration at different topographic positions and catchment water balance following a pronounced drought in a mixed species eucalypt forest, Australia
Highlights
► Controls on spatial and temporal variation in forest water budget after drought were quantified. ► Evapotranspiration was consistent between years, driven by evaporative demand and landscape position. ► Soil water deficits decreased stream flow for several years following a severe drought. ► Drought decoupled evapotranspiration and soil water from stream flow due to changes in storage.
Introduction
Drought and widespread water shortages across Australia are highlighting the vulnerability of Australia’s water resources to climate change and catastrophic events such as large-scale wildfires (Lane et al., 2010, Petrone et al., 2010). Across southern Australia, a large proportion of the urban water supply is sourced from mountainous catchments forested with native eucalypts. Like many watersheds across the world, the impacts of climate change in south-eastern Australia are likely to reduce the supply of water, via changes in precipitation (P) patterns (Chiew et al., 2008) and increases in atmospheric demand, via increases in mean annual temperature (Howe et al., 2005). Water yield in this region is highly dependent on how different forest functional types affect energy and water exchange across a catchment in response to changing climate and disturbance patterns.
The mixed species eucalypt forest (MSEF) is the most common forest type in the montane regions of south-eastern Australia and is functionally and structurally distinct from the adjacent, wetter ‘ash’ type forests (Eucalyptus regnans F. Muell. and Eucalyptus delegatensis R.T. Baker). This forest type occurs predominately at low-elevations and includes the silvertop ash (Eucalyptus sieberi L.A.S. Johnson) – stringybark species group and the messmate (Eucalyptus obliqua L’Hér) – peppermint (Eucalyptus radiata Sieber ex DC. and Eucalyptus dives Schauer) species group (Lutze et al., 2004). The MSEF generally occurs at sites with mean annual P < 1300 mm (Kellas and Hateley, 1991) but inter-annual variation in P can be >50% (Lane and Mackay, 2001, Roberts et al., 2001, Bren and Hopmans, 2007). These forests occupy drier and more exposed habitats that experience greater fire frequency and are generally classified as being ‘fire tolerant’ because they regenerate via resprouting from stems, lignotubers and roots or from seed (Gill and Catling, 2002).
The hydrologic dynamics of wetter eucalypt forest types, such as mountain ash (E. regnans) forests have been studied intensively (Kuczera, 1987, Watson, 1999, Vertessy et al., 2001) but for many of the drier forest types such as the MSEF, factors controlling forest water use and stream flow (Qstream) are poorly understood. The inter-annual variability of the rainfall–runoff relationship in southern Australia is high, particularly at low annual P (Zhang et al., 2001, Peel et al., 2002). Furthermore, empirical meta-analysis of Australian catchments showed that catchment water balance was highly sensitive to vegetation dynamics, soil type and topographic properties where climate dryness indices [ratio of potential evapotranspiration (Epot) to P] are close to 1 (Zhang, 2004), suggesting Et becomes difficult to predict without more detailed information on these catchment and climate properties. These broad generalisations seem to be consistent with catchment monitoring in the MSEF in north-eastern Victoria, where annual Qstream can fluctuate widely (0–214 mm) during years where P < 1150 mm (mean annual Epot/P = 0.85) (Bren and Hopmans, 2007). We expect the catchment water balance of MSEF to be particularly responsive to warmer and drier conditions expected under climate change, because Qstream represents a smaller proportion of annual P (Bren and Papworth, 1991, Lane and Mackay, 2001, Bren and Hopmans, 2007) and disturbances that alter forest structure such as wildfire may become more frequent.
The high inter-annual variability of catchment water yield characteristic of the MSEF and other similar forest types highlights the need to understand the spatial and temporal controls on forest water use and the effect of antecedent soil water conditions on Et and Qstream. Spatial heterogeneity in soil moisture and Et is rarely assessed for catchment water budget studies and might be critical to the Qstream signal as the source for water entering the stream changes in response to the amount of water in the soil and its transport potential (matric and gravitational influences on water movement). In southern Australia, where rainfall is either winter dominant or uniformly distributed through the year, Qstream is often out of phase with Et, meaning that rainfall must replenish soil water lost during summer before Qstream can recover in winter. These patterns affect the connectedness between horizontal and vertical fluxes of water in the catchment and can lead to a decoupling between Et, soil and stream fluxes (Brooks et al., 2010) that manifest as time lags in Qstream recovery at increasingly longer timescales (Moore et al., 2011) and may lead to a temporal mismatch in fluxes measured at different temporal scales.
For most forest types, including southern Australia, data on the various catchment water balance components are relatively few. Those available from native eucalypt forests are for single stands (Mitchell et al., 2009) and confound the effect of spatial variation (Loranty et al., 2008) and stand age (Vertessy et al., 1995, Roberts et al., 2001). Water use by trees is usually the largest flux in native eucalypt forests and its response to variability in P is likely to be related to the extent to which energy or soil water limits stand water use. For example, tree water use is tightly controlled in water limited environments and shows some plasticity in response to inter-annual variation in P. For example, Zeppel et al. (2006) showed that Esap remained at a constant fraction of P over 2 years of different annual P in an open Eucalyptus and Callitris woodland. Less inter-annual variability in forest water use was observed in a wet sclerophyll forest where annual totals of Et (measured using eddy covariance) during drought and non-drought years were relatively similar due to the trees having access to deep soil water (Leuning et al., 2005). The contribution of additional components of the water balance, interception (Ei) and understorey/forest floor evapotranspiration (Efloor), can be significant given the low intensity rainfall patterns characteristic of southern Australia (Pook et al., 1991, Haydon et al., 1996). Ei occurs during and directly after the rainfall event, so it responds at shorter timescales to changing rainfall patterns, atmospheric demand and/or variation in forest type and structure (Crockford and Richardson, 1990). Clearly, understanding the response of Et to water and energy limitation over seasonal and annual timescales in different forest types will enhance our ability to predict water yield responses under a range of climatic and disturbance scenarios.
This study investigated the water balance components of a relatively undisturbed, low-elevation MSEF catchment using a ‘bottom up’ approach involving field measurements of Esap, Ei, Efloor and Qstream. The study aimed to assess temporal and spatial variation in Et by (1) quantifying the contribution of the different Et components over seasonal and annual periods, (2) assessing the variation in Et across the catchment arising from topographic position and (3) evaluating the role of water supply and demand limitations on the catchment water balance in the context of climate variability and drought.
Section snippets
Study site
This study was conducted at Long Corner Creek (LCC) catchment approximately 250 km north east of Melbourne (36°41′S, 146°39′E). The LCC catchment is made up of two sub-catchments LCC A (133 ha) and B (87 ha) which are tributaries to the Buffalo River. All measurements reported in this paper were obtained from LCC A. The catchment lies in the middle of the Ovens River catchment, an important tributary of the River Murray and the majority of which is vegetated by native forest (53.5%) and plantation
Meteorological, soil water and stream flow dynamics
Mean daily D ranged from a mean of 1.9 KPa during the summer months to 0.2 KPa during winter (Fig. 2a). Daytime totals of Rn reached a maximum during December (21.6 MJ m−2 d−1) and were minimal in June (−1.3 MJ m−2 d−1, Fig. 2b). Daily rates of E0 ranged from 0 to 8.7 mm d−1 and remained below ∼2 mm d−1 for 4 months between April and August (Fig. 2c). Annual P during the Et monitoring period for the 2008–2009 and 2009–2010 water year (May–April) were 869 and 1196 mm respectively and on average 60% of rain
Factors controlling forest water balance
The seasonal and annual patterns of Et from both the observed and the modelled data show that Et is largely driven by changes in energy availability. In dry years, the trees appeared to have access to a substantial store of deep soil water that enabled annual Et to be maintained within a relatively narrow band compared with inter-annual variation in precipitation (Fig. 10b). Meta-analyses of catchment behaviour have suggested that P is the most dominant factor in determining Et for forested
Conclusions
This study presents the first detailed analysis of the catchment water balance components in the MSEF and defines the response and recovery of water yield during drought. The 40% spatial variation in annual Et within the catchment appeared to be associated with landscape position and changes in stand structure and water use scalars – LAI and sapwood area. Esap was the most dominant flux annually and we showed that at the individual tree-level, water use patterns could be by a single
Acknowledgements
The authors would like to thank John Costenaro, John Collopy, Philip Noske and Chris Sherwin for their support in the field. The Department of Sustainability and Environment Victoria are acknowledged for providing permission to work at the LCC catchment and road access to the monitoring sites. This study was funded through a CSIRO Water for a Healthy Country Flagship collaboration grant.
References (60)
- et al.
Paired catchment observations of the water yield of mature eucalypt and immature radiata pine plantations in Victoria, Australia
J. Hydrol.
(2007) - et al.
Using spatial interpolation to construct a comprehensive archive of Australian climate data
Environ. Model. Softw.
(2001) - et al.
Modelling historical lake levels and recent climate change at three closed lakes, Western Victoria, Australia (c.1840–1990)
J. Hydrol.
(2001) Prediction of water yield reductions following a bushfire in ash-mixed species eucalypt forest
J. Hydrol.
(1987)- et al.
Modelling the long term water yield impact of wildfire and other forest disturbance in Eucalypt forests
Environ. Model. Softw.
(2010) - et al.
Streamflow response of mixed-species eucalypt forests to patch cutting and thinning treatments
For. Ecol. Manage.
(2001) - et al.
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.
(2005) Estimation of leaf area index in a eucalypt forest using digital photography
Agric. For. Meteorol.
(2007)- et al.
Partitioning of evapotranspiration in a semi-arid eucalypt woodland in south-western Australia
Agric. For. Meteorol.
(2009) - et al.
Transpiration from Eucalyptus sieberi (L. Johnson) forests of different age
For. Ecol. Manage.
(2001)