Cap and trade policy for managing water competition from potential future carbon plantations

Environmental Science & Policy 66 (2016) 11–22

Contents lists available at ScienceDirect

Environmental Science & Policy journal homepage: www.elsevier.com/locate/envsci

Cap and trade policy for managing water competition from potential future carbon plantations Jeffery D. Connor* , Brett A. Bryan, Martin Nolan CSIRO Waite Campus, Urrbrae, South Australia

A R T I C L E I N F O

Article history: Received 31 March 2016 Accepted 10 July 2016 Available online xxx Keywords: Water trade Carbon sequestration Integrated assessment Forest water interception Land use economics Reforestation

A B S T R A C T

Carbon sequestration from reforestation can play a large role in mitigating global climate change. However, resulting interception of rainfall runoff may impose high irrigation, water supply and/or environmental flow costs. This article presents an assessment of water trade policy to manage fresh water supply, carbon sequestration trade-offs for the Murray-Darling Basin. A linked Australian high spatial resolution land use and global integrated assessment framework evaluated plausible and internallyconsistent global scenarios to 2050 involving significant carbon planting incentive. Substantial flow loss from increased interception was estimated absent policy to balance carbon water trade-offs. Absent policy to address the trade-off, irrigation opportunity costs was estimated to substantially exceed carbon sequestration economic value in futures with significant carbon sequestration incentive. The value of integrating interception from new carbon plantings into the existing water trade system was estimated at $3.3 billion and $2.0 billion (2050 annual value) for our strong and moderately strong global climate action outlooks with our reference case assumptions. The conclusion that trade provision in policy to cap interception impacts can produce significant benefits in scenarios with significant carbon sequestration incentive remained robust over a very broad set of sensitivities tested with benefit estimated at over $1 billion annually at 2050 even for very conservative assumptions. Crown Copyright ã 2016 Published by Elsevier Ltd. All rights reserved.

1. Introduction Significant potential exists for carbon emissions abatement through incentive to encourage reforestation of agricultural land (Bryan et al., 2014; Benitez and Obersteiner, 2006; Lubowski et al., 2006). However, significantly reduced water supply for human uses and the environment can result (Egginton et al., 2014; Bryan et al., 2015a). This is because forested land intercepts and evapotranspires more water than land covered with crops, shrubs, or pasture (Brown et al., 2007; Chu et al., 2010; Farley et al., 2005). Carbon-water trade-offs resulting from reforestation incentive are most challenging and most evaluated in semi-arid regions where greater proportional runoff reductions result compared with higher rainfall environments (Jackson et al., 2005). An example is the 872,000 ha of forest establishment estimated to reduce water availability by 8% in the southern Murray-Darling Basin in response to a $100/t carbon price (Schrobback et al., 2011). Similar trade-offs have been identified in the Fynbos ecoregion, South Africa (Chisholm, 2010), in individual sub-catchments within the

* Corresponding author. E-mail address: [email protected] (J.D. Connor). http://dx.doi.org/10.1016/j.envsci.2016.07.005 1462-9011/Crown Copyright ã 2016 Published by Elsevier Ltd. All rights reserved.

Murray-Darling Basin (Bathgate et al., 2009; Nordblom et al., 2010; Nordblom et al., 2012), and in semi-arid parts of western China where major afforestation has taken place over the past decade (Gao et al., 2014). Carbon-water trade-offs from reforestation can also arise in higher rainfall areas such as New Zealand in locations where intercepted runoff would otherwise provide irrigation water supply (Dymond et al., 2012). Despite recognition of the need for policy to manage carbon forest competition for water (Egginton et al., 2014; Calder, 2007; Young and Mccoll, 2009), actual policy to address the issue has been limited to date and implemented primarily through land use regulation. For example, in South Africa, new plantation forests require a permits which are only approved after investigation and agreement by relevant State agencies that stream flow impacts are acceptably small (Kruger et al., 2008). Zoning of where carbon forest incentive policy is allowable is another land use regulation approach. For example, the latest Australian carbon farming incentive policy defines zones where historic average rainfall exceeds 600 mm per annum as not eligible for carbon sequestration incentive payments because of the potential for significant run-off interception (Australian Government, 2015). A challenge with regulatory approaches like those implemented in South Africa

12

J.D. Connor et al. / Environmental Science & Policy 66 (2016) 11–22

is the time, effort and cost of the bespoke individual investigation and approval requirements. This can lead to very limited land use change, even where it may be possible without unacceptable negative consequences (Kruger et al., 2008). Zoning approaches like those implemented in the Australian carbon farming initiative overcome this transaction cost impediment. However, these are blunt instruments which coarsely target high interception impact land. In some cases reforestation of significant areas within the 600 mm per annum rainfall zone may have little consequential impact on runoff (Van Dijk et al., 2007). Cap and trade policy could be a more efficient approach to address water trade-offs arising from carbon reforestation incentives. The approach has been effectively used to manage common pool resources including sulphur dioxide emissions (Schmalensee et al., 1998), carbon emissions (Paltsev et al., 2008; Grubb, 2012), and water diversions in several western US states (Lane-Miller et al., 2013), Spain (Kahil et al., 2015), Chile (Hearne and Easter, 1997), and in the Murray-Darling Basin (Kirby et al., 2014; Grafton and Horne, 2014). Applied to forest water interception, the approach would effectively cap the total effect on river flow of withdrawals for irrigation and interception from reforestation. New forests would need to compete for water via the purchase of water rights from regional irrigators who draw on the same water resources that would be impacted (NWC, 2011). The approach has the advantage that it could maintain a balance

between benefits of climate mitigation from carbon sequestration and resulting water supply opportunity costs in a way that dynamically adjusts to evolving and uncertain land and water supply and demand drivers (Young and Mccoll, 2009). In this study, we assessed potential for increasing carbon planting area to reduce runoff and river flows for the MurrayDarling Basin, Australia. We considered two global outlooks with growing carbon price and continuing Australian national policy incentive for carbon sequestration and three policy approaches to address carbon planting water interception. A no cap scenario considered a future without any attempt to limit flow losses from carbon planting water interception. Two cap scenarios evaluated policies to address reduced flow from additional interception. One cap policy scenario included the flexibility of a trading mechanism: this required landholders changing from current agricultural land uses to carbon plantings to purchase water rights from current irrigation water rights holders for the water intercepted. The other cap policy scenario included no water trading: flow balance was maintained by reducing irrigation water supply by the amount of growth in interception from new carbon plantings. We assessed the potential impacts of the three policy scenarios on land use, water use, carbon sequestration, and economic returns; mapped the spatial patterns of land use change; and undertook sensitivity analysis of the outcomes to variations in key parameter driving outcomes.

Fig. 1. Murray-Darling Basin study region.

J.D. Connor et al. / Environmental Science & Policy 66 (2016) 11–22

13

2. Methods

2.2. Modelling approach

2.1. Case study

Our modelling platform, the Australian Land Use Trade-offs (LUTO) model (Connor et al., 2015; Bryan et al., 2014), has been extensively applied (Grundy et al., 2016; Bryan et al., 2015b) and evaluated (Gao et al., 2015). It estimates long run land use change decisions as they evolve on annual time steps from 2013 to 2050 and impacts on food, carbon, water, energy, and biodiversity ecosystem services. Modelling is at
1.1 km grid cell resolution within the context of internally-consistent outlooks for the longterm evolution of global climate, population, technology, greenhouse gas emissions, carbon and energy prices, and food demand (Hatfield-Dodds et al., 2015b). Here, we focused on two global outlooks to 2050. As described in detail in Hatfield-Dodds et al. (2015a), the L1 outlook involved very strong global greenhouse gas abatement leading to a comparatively high global carbon price of around $200 tCO21 by 2050 and comparatively low CO2 concentration growth. The M3 outlook involved somewhat greater population growth, and strong global greenhouse gas abatement effort with global carbon price growth to $118 tCO21 by 2050 (Fig. 2). Carbon, oil, crop and livestock price trajectories from the outlooks are shown below (Fig. 2). To model a future with significant carbon forest area growth, we assumed continuation of current Australian policy which provides financial payments for offsetting of greenhouse gas emissions through reforestation of cleared agricultural land (Australian Government, 2015). Land use possibilities considered in this analysis include the current agricultural land use (classified as one of 24 irrigated or rain-fed agricultural commodity options) and carbon plantings (Eucalyptus monocultures). In other applications we also consider options of changing to forest land use for biofuel or bioenergy feedstock production or biodiversity (Grundy et al., 2016), but these were precluded here to focus on carbon planting carbon sequestration and water interception impacts. Eqs. (1)–(6) describe the mathematical programming model solved to determine land use change each year in this study. Baseline year land use, irrigation water diversion and economic returns are the starting point. From this point returns to agricultural and carbon planting land uses are updated with

Our study area is the 35.2 Mha of the 84 Mha Murray-Darling Basin (MDB) that are presently cleared of forest for irrigated and dryland agricultural production (Fig. 1). The MDB is Australia’s most significant agricultural region providing more than 40% of dryland and 60% of irrigated agricultural production nationally (Kirby et al., 2014) and includes some of Australia’s most productive forestry sites (Polglase et al., 2008). Both irrigation and environmental assets such as the Ramsar-listed Coorong river delta estuarine lake and wetland region could be at significant risk if increased carbon forest water interception reduced MDB flows. Another reason for focus on the MDB is that incorporating water interception into the already existing and arguably world’s most developed water diversion cap and trade scheme would require less institutional development effort than in jurisdictions with less effective existing tradeable property right systems (Grafton et al., 2016). The Australian National Water Initiative (NWI), a Commonwealth program to monitor and set practice standards for State and regional water policy, endorsed incorporating water interception into a framework that would facilitate trading interception with other water uses. The NWI recommended that States move toward policy where “existing significant interception activities in water systems that are fully allocated, over-allocated, or approaching full allocation would be recorded, and that new activities would require a water access entitlement” (NWC, 2011). South Australia is the only state to act on the recommendation to date with 2013 legislation declaring plantation forests to be a “water affecting activity” requiring a water entitlement in specified regions and issuing water entitlements to existing forest plantations in the south-east region, home to most of the State’s plantation forests (Gillet et al., 2014). Conceptually, these entitlements are tradeable with other water extraction entitlements belonging to regional irrigators who draw on the same water resource that is impacted by forestry (NWC, 2011), though no trade has taken place to date (Gillet et al., 2014).

Fig. 2. Global integrated assessment modelled crop, livestock, oil, and carbon price paths under the two global outlooks.

14

J.D. Connor et al. / Environmental Science & Policy 66 (2016) 11–22

revised agricultural commodity demand, and carbon and input prices that are produced externally to LUTO with a global integrated assessment climate and emissions plus world economy computable general equilibrium model (Newth et al., 2015). An optimisation model is solved to estimate levels and locations of long run permanent land use change that occur each year over the modelling horizon from 2013 to 2050 given the evolving annual changes in prices and demands. One innovative feature is the possibility to model alternative assumptions about how competition for agricultural land could influences food price. The endogenous price formulation allows accounting for how feedback through greater agricultural returns can moderate rates of land use change to carbon plantings. This formulation is appropriate for the assumption that the scale of land use change response to carbon price incentives globally is proportionally similar to the Australian land use change response. The possibility to model such food price feedbacks is important given the evidence from previous related studies demonstrating that such feedback can be a significant determinant of land use change and food supply in scenarios where changes have significant global supply impacts (Dumortier et al., 2011; Stavins, 1999; Golub et al., 2013). Functionally, this is represented with a classical partial equilibrium formulation involving Eqs. (1) and (2) in combination with Eqs. (3) and (5) and explained in more detail in Connor et al. (2015). In Eq. (1), the objective is to maximise the sum of consumer surplus (the first term in Eq. (1)) and producer surplus (the second term in Eq. (1)). Agricultural commodity prices Pi in Eq. (1) are determined endogenously with the inverse price quantity relationships in Eq. (5). The term ei in Eq. (5) is the price elasticity of demand by food commodity i which is parameterised with values sourced from Andreyeva et al. (2010), and with baseline demands calibrated to observed demands at prices observed in the baseline year. Eq. (3) requires that the sum of supply from all land use cells in solution to at least equal demand for each agricultural commodity i. Eq. (2) represents returns to carbon plantings where the price of carbon Pj increases with each annual iteration on a trajectory determined exogenously to LUTO with our global integrated assessment model (Newth et al., 2015). In the food price endogenous scenario, as the model iterates, the demand curves for agricultural commodities shift outward on global demand growth trajectories determined by the global integrated assessment model, and rising competition for land for carbon plantings reduces supply of agricultural commodities available at any given price level. The confluence of growing demand for agricultural commodities and less potential to supply at any given price drives up equilibrium agricultural commodity price relative to what would result absent this food price feedback. This endogenous price formulation of LUTO is consistent with assuming carbon sequestration policy and food supply impacts from agricultural land conversion to carbon plantings are proportionally similar in Australia and globally. We also implemented the model with a price exogenous formulation which is consistent with a scenario where Australia acts unilaterally or as one of very few nations providing carbon sequestration incentive. In this case no food price feedback is assumed because reductions in global supply of land for food from carbon plantings is assumed to have such a small influence on global supply that it is reasonable to ignore the negligible resulting global food price impacts. In this formulation, Eq. (5) is omitted and agricultural commodity prices are determined exogenously in the global integrated assessment model and passed into LUTO at each annual iteration without modification. Another unique feature of LUTO is the inclusion in Eq. (2) of an adoption hurdle rate, h, requiring new carbon planting land use to be h times as profitable as current agricultural land use for land use change to occur. It represents how given inherent risks of long term

commitments and high upfront cost, landholders typically only commit to land use change involving reforestation when returns exceed current agricultural land use returns by a considerable margin. Several lines of evidence are all consistent with our base case assumption hurdle rate value of h = 2. This includes findings from real options analysis of triggers required to induce change from farming to trees (Regan et al., 2015), comparison of profit function and econometrically derived agricultural to forest land use change estimates (Richards and Stokes, 2004), and revealed discount rates in actual decisions to convert from agriculture to timber (Prestemon and Wear, 2000). In sensitivity analysis we also assessed h = 1 an assumption consistent with classic expected value profit maximisation assumptions and an h = 5 assumption more representative of conversion reticence in highly uncertain policy environments. See online Supplemental material for more on this assumption. Maximize: X ðPi  C ir Þ  qir air ð1Þ i;r

þ

X 1=h  ðPj  C jr Þ  qjr ajr

ð2Þ

j;r

subject to: X qir air  yi 8i 2 I r

X air þ ajr ¼ 1

8r 2 R

i

1=ðei Þ

Pi ¼ ai yi

8i 2 I

XX ½ wir akr þ wjr ajr þ  WA r

ð3Þ

ð4Þ

ð5Þ

ð6Þ

i

Indices: r = grid cells i 2 I = 24 agricultural commodity land uses, j = carbon planting land use, Variables: qir, qjr, = production quantity of agricultural (i), carbon planting (j) from cell r yi = quantity of agricultural commodity i produced in segment s2S Parameters: Pi, Pj = price of agricultural commodity i,carbon j Cir, Cjr, = production cost of agricultural commodity (i), carbon planting (j) for cell r air, ajr = binary indicator variable if optimal land use choice in cell r h = hurdle multiplier representing rate by which carbon planting land use return must exceed current agricultural land use return for land use to change wir = irrigation water use for agricultural commodity i for cell r, equals zero for dryland agricultural activities wjr = reduced flow from forest water interception from carbon planting land use j for cell r Eq. (6) is the water use constraint applied for modelling cap and trade policy. This requires water use for irrigated agriculture land use, wir plus water intercepted by conversions to carbon plantings, wjr to be less than or equal to total water availability in year t, WAt.

J.D. Connor et al. / Environmental Science & Policy 66 (2016) 11–22

This formulation is a common way to represent water trading in mathematical programming models (e.g. Qureshi et al., 2007; Kahil et al., 2015). This constraint works by requiring levels of activities competing for limited water (in this case land use pixels using water for irrigation or interception offset) to adjust until marginal values of the last units of water used in competing activities are equilibrated. In the no cap scenario water interception from carbon planting land use is exempt from the water constraint with wjr set equal to zero for all cells r in Eq. (6). In the cap without trade scenario, we assume that increases in forest water interception don’t require water rights but do lead to reduced irrigation water allocations. This is represented by setting wjr equal to zero in Eq. (6) but calculating water interception in each period t and subtracting this amount from water available for irrigation in the next period t + 1. 2.3. Data Carbon planting tree growth and carbon sequestration estimates for Australia were modelled based on fast-growing, locallyadapted Eucalyptus species with growth influenced by climate and soil variables (Polglase et al., 2008). Climate change, fire, and drought risks were considered (Bryan et al., 2014) and sequestration estimates were further discounted by 20% to cover residual risks and uncertainties. Spatially-differentiated estimates of water impacts of reforestation varying at 0.05  resolution depending on rainfall and soils were taken from the Australian Water Resources Assessment—Landscape model (Van Dijk and Renzullo, 2011). This data characterises expected water run-off (ML ha1 yr1) changes from land use conversion to Eucalyptus forest from annual crops and pastures. Spatially differentiated irrigation water use (ML ha1 yr1) estimates by crop were sourced from (Marinoni et al., 2012). Returns to agricultural commodities were calculated using a profit function approach (Marinoni et al., 2012) for each simulation year this was calculated as revenue (price yield) less fixed and variable costs of production with spatial differentiation in yields and costs based on observed base year variability (Marinoni et al., 2012). Under each global outlook, for each year, returns were updated by the impact of climate change, assumed rates of yield productivity increase, the impact of global food demand on commodity price, and the impact of global oil price on production costs. Annual carbon planting payment value was calculated as carbon sequestration times global integrated assessment model determined carbon price in each simulation year assuming a 100 year economic asset life. To represent the opportunity cost of permanent land use change agricultural returns were also expressed as net present values over a 100 year horizon. In both cases, we used a 10% p.a. discount rate representative of the long run average commercial rate on credit faced by Australian farmers over the previous three decade time horizon. 2.4. Policy scenarios 2.4.1. No cap scenario This scenario assessed implications of agricultural conversion to carbon plantations without any policy to limit resulting water interception. Carbon planting area in response to rising carbon price increased water interception with no irrigation water extraction level adjustment to protect river flows. While the value of riparian environmental asset damage that could result in such a scenario from reduced flow is significant, we quantified only the volume of flow reduction. 2.4.2. Cap with and without trade scenarios These scenarios assessed impacts of two approaches to regulating water interception impacts of carbon planting land use expansion. Both capped the sum of irrigation water extraction

15

plus growth in interception at baseline (2006/7) irrigation diversion less an annual subtraction equal to reduced inflow from climate change (a 6% cumulative decline to 2050 for the L1 outlook and 4% for the M3 outlook) as modelled by Grigg et al. (2013). In the cap without trade scenario, irrigation water allocation was reduced in each annual time step by the amount of carbon planting water interception increase in the previous period. In addition, cumulative interception from carbon plantings was constrained to be less than or equal to total baseline irrigation diversions, allowing no additional carbon plantings if this limit was reached. The cap and trade scenario also involved constraining the sum of irrigation water extraction plus growth in interception at the same level as in the cap without trade scenario. However, to change from agriculture to carbon plantings, landholders were required to attain water rights to offset interception flow reduction through purchase of an equal amount of irrigation water diversion entitlement in the water market. We assumed two distinct water markets to represent the lack of hydrological connectivity in most years between the southern MDB (including the Murray and Murrumbidgee Rivers and tributaries) and the northern MDB (the Darling River and tributaries). Caps were applied to the northern and southern basin separately and trading water between the two sub-basins was precluded. 2.5. Sensitivity analysis We assessed sensitivity of estimated economic benefits of cap and trade policy to three key assumptions identified as highly influential and in previous systematic sensitivity analysis of LUTO model outcomes (Gao et al., 2015; Gao and Bryan, 2016) and in previous related studies. Uncertain future improvements in agricultural yields, a key outcome driver in related land use change studies (Audsley et al., 2015; Kriegler et al., 2012) was tested with a default growth rate of 1.5% per annum (close to the Australian historical average) as well as a higher 3% rate consistent with optimistic productivity growth assumptions and a lower 0.5% rate consistent with the slow Australian productivity growth observed from 1997 to 2007 (Hughes et al., 2011). Uncertainty in the degree of uptake of new land uses with high upfront investment requirements, delayed and uncertain returns modelled with our default assumption that conversion from agriculture to carbon planting land use only occurred when return to new carbon planting land use was double the return to current agricultural land use (h = 2). We also tested sensitivity with alternative assumptions that carbon planting returns five times greater (h = 5) and equal to (h = 1) current agricultural land use returns would motivate conversion. The impact of land competition on food prices was the third key uncertainty tested. The reference case assumption was high food price feedback sensitivity modelled by assuming global reductions in food supply from agricultural land conversion to carbon plantings were proportionally similar to the Australian food supply reductions that LUTO estimates modelled with the food price endogenous formulation of LUTO described above. The alternative assumption tested was no food price feedback consistent with an assumption that reductions in global supply of food from conversion of Australian land to carbon plantings represents have very small impact on global supply and price. More detail on sensitivity analysis implementation and parameter value assumption data basis is discussed in Supplemental online material. 3. Results 3.1. No cap policy Under the no cap policy, significant land use change, water, and carbon impacts were projected for both moderate (M3) and large

16

J.D. Connor et al. / Environmental Science & Policy 66 (2016) 11–22

Fig. 3. Carbon plantings, irrigated and dryland agriculture water use, carbon sequestration and economic returns by policy scenario for L1 outlook (top) and M3 outlook (bottom).

J.D. Connor et al. / Environmental Science & Policy 66 (2016) 11–22

(L1) carbon price rise outlooks. The L1 outlook, with carbon price rising to $200 per tonne by 2050, led to 22.27 million ha of carbon plantings replacing dryland agriculture and some irrigation by 2050 (3a top frame) and sequestration of 250.57 million tonnes of CO2 per year by 2050 (3 g top frame). An adverse consequence was 8370 GL in flow losses from carbon forest interception by 2050 (3d top frame): an amount exceeding baseline (2006/7) irrigation diversion. Outlook M3 land use, water, and carbon impacts were qualitatively similar to L1 outlook estimates for the no cap scenario (Fig. 3 leftmost column, bottom frame) but of lesser magnitude. In this outlook by 2050, carbon plantings grew by 13.67 million ha (3a bottom frame), carbon sequestration grew to 188.09 million tonnes annually (3 g), with a cost of 5160 GL of intercepted basin inflow (3d)—a volume equal to around 79% of baseline irrigation diversion.

17

typically experienced in implementation of cap and trade policies whereby knowledge that a resource will soon be capped leads to increased use of the resource just prior to capping (Hanemann, 2010). Of course, the policies capping water use come at some cost, most notably significantly less carbon sequestration. Economically, the cost is expressed as reduced returns from carbon plantings and irrigated agriculture. These losses are somewhat offset by increased returns to dryland agriculture. As discussed in greater detail in the next subsection, the cap and trade policy allowed the same level of flow protection at less cost than the cap without trade approach—a net $A 0.5 B cost with trade as opposed to $A 2.52 B cost without trade annually at 2050 and for the M3 outlook a $A 0.5 B with trade as opposed $A 3.9 B without trade 2050 annual cost for the L1 outlook (Fig. 4). 3.3. Common trends across policies

3.2. Impacts of capping water resource use Carbon planting area, sequestration, and water interception differed significantly between policy scenarios. Magnitudes were dampened via implementation of cap policies relative to the no cap policy because of the limits and costs on water interception that constrained opportunities to expand carbon plantings. Fig. 4 summarizes the 2050 estimated results from cap scenarios with and without trade relative to the no cap scenario. Both cap policies mostly precluded growth in water use. The small increases observed for the two cap policies resulted from time lag effects of adjusting irrigation diversion limits down by the amount of water interception in the previous period. While this outcome is not fully consistent with the intent of completely precluding loss of flow, it is consistent with “announcement effect” outcomes

Several trends were shared across all three policy scenarios (no cap, cap without trade, and cap with trade) driven by shared L1 and M3 global outlook driver trends. This included trends driven primarily by growing carbon price trajectories including increasing: carbon planting area (3a–c), carbon sequestration (3g–i), water interception (3d–f). Area in dryland agriculture also declined for all three scenarios with increasing conversion from this land use to carbon plantings. Another commonality across all scenarios was growth in economic returns to all land use (3j–l top and bottom frames) over time. This resulted from common global outlook driver trends of growing agricultural commodity prices, carbon price, and agricultural productivity. This even held true for irrigated agriculture despite less irrigation water diversion in the two cap scenarios (3e & f top and bottom frames) and for dryland

Fig. 4. Economic and water impacts of capping carbon planting water interception impacts with and without water trade.

18

J.D. Connor et al. / Environmental Science & Policy 66 (2016) 11–22

agriculture despite large reductions in area in this land use. For dryland and irrigated agriculture, food price and agricultural productivity growth effects outweighed impacts of less water for irrigation and less land for both dryland and irrigated agriculture. 3.4. Value of water trading provisions Estimated 2050 annual returns to land under the cap and trade policy were $3.3 B greater than under the cap without trade policy for the L1 outlook and $2.0 B greater for the M3 outlook (Fig. 5). This primarily resulted from avoided land use conversion to carbon sequestration at locations where flow loss value would have exceeded forgone irrigation opportunity cost. The trade scenario resulted in greater area, water use, and economic returns for irrigated agriculture (Fig. 5). Higher returns to carbon plantings were also found, despite less cumulative carbon sequestration for the cap and trade, as opposed to the cap without trade scenario. This somewhat counterintuitive finding resulted from our application of the cap without trade policy where further carbon plantings were precluded if total interception from cumulative growth in carbon plantings reached total irrigation diversion in either the northern or the southern sub-basin. This limit was reached the northern sub-basin and more quickly in the cap without trade scenario, and consequently a greater proportion of total area of carbon plantings occurred in later years at higher prices in the cap with trade scenario. 3.5. Cap and trade policy benefit and sensitivity Sensitivity analysis results revealed significantly greater policy benefit for higher agricultural productivity rates, higher carbon price growth, and lower hurdle rates required for land use conversion. Greater policy benefit was also estimated for the assumption of no price feedback from land competition (exogenous price assumption) versus our default endogenous price

modelling approach consistent with an assumption that significant global land use conversion to carbon plantings proportionally similar to what is estimated for the MDB increases food price through supply and demand feedbacks. The reason for the greater benefit in all cases was that the alternate assumptions created greater incentive for carbon plantings and thus greater benefit from reduced water interception at high irrigation opportunity cost. One interpretation of sensitivity analysis findings is that our analysis provides conservatinve estimates of cap and trade policy value in contexts of significant carbon sequestration incentive. This is because key default assumptions can be interpreted as conservative. Most notably the assumed requirement for double the returns to carbon forest to motivate land use change and the assumption that carbon forest land competition significantly increases food price and dampens land use conversion significantly reduce estimated policy benefit value. A number of the alternate assumptions that could be reasonably justified such as land use conversion at equal carbon forest and agricultural returns and little food price feedback from carbon forest land competition are consistent with the realisation of significantly higher benefit from trade provision in cap policy. An interpretation of the small estimates of trade policy value for some sets of assumptions, particularly high hurdle rate assumptions, could be that potential to realise policy benefit is particularly sensitive to contexts where policy uncertainty means that landholders require very high rates of return relative to earnings from current land uses to switch to carbon plantings. Other than in this context our finding of high potential for significant policy benefit is robust: under nearly every other set of assumptions tested the annual value of trade provision benefits in the cap policy was estimated to exceed $1 billion annually by 2050 (Fig. 6). 3.6. Spatial variation in land use For the no cap policy, with more moderate carbon price trajectories in the M3 scenario, most carbon planting occurred in the relatively low net return extensive dryland livestock grazing land in the north and east of the study area. While the higher carbon price L1 outlook also included carbon plantings on higher opportunity cost and southern basin dryland cropping and livestock land. The mapping also shows that the two cap policies led to significant reductions in carbon planting area overall relative to the no cap scenario. Comparing the cap with and without trade maps shows a shift in carbon planting location away from highest rainfall areas along the Great Dividing Range (inland from the east coast of Australia), and from the most northern part of the basin to lower rainfall areas with less water interception opportunity cost in the cap with trade scenario. More irrigation in the cap with trade as opposed to without trade scenario is also visible on the maps (Fig. 7). 4. Discussion 4.1. Analysis scope and innovation

Fig. 5. Economic and environmental impacts of water trade provisions in policy to cap water resource use (i.e. cap with trade compared to cap without trade).

This article provided assessment of water policy options for managing water resource impacts of increasing carbon forest land use possible under global change scenarios. The work extends previous Australian land use scenario modelling (Grundy et al., 2016) by evaluating potential to integrate interception flow losses into an existing water diversion cap and trade policy that is already producing large benefit (Grafton and Horne, 2014). A key advance on previous evaluation of the issue is deployment of global best practice in land use change evaluation with a high resolution national land use change model linked with integrated assessment

J.D. Connor et al. / Environmental Science & Policy 66 (2016) 11–22

modelling of global, plausible, internally consistent climate, policy, economic and resource scenarios to 2050. The analysis also advanced in the treatment of uncertainties previously limited to consideration of carbon price in isolation (e.g. Nordblom et al., 2012; Schrobback et al., 2011) with sensitivity analysis for key factors shown to be particularly significant outcome drivers in previous related carbon and biofuels land use policy evaluations including uncertainty in future improvements in agricultural yields (Audsley et al., 2015; Kriegler et al., 2012), in the degree of uptake of new land uses in response to incentive (Schmitz et al., 2012; Ewert et al., 2005), and in the impacts of land competition on food prices (Dumortier et al., 2011; Stavins, 1999; Golub et al., 2013). 4.2. Key findings Results showed potential for large Murray-Darling Basin flow losses. Estimated water interception from reforestation exceeded current irrigation diversion by 2050 for the most extreme future scenario with: carbon prices rising to $200 CO2 t1, continuing Australian policy incentive for carbon sequestration, and no policy to limit resulting water interception impacts. A cap without trade policy reduced the amount of carbon planting area and resulting interception relative to the no cap scenario but at a significant cost of reduced irrigation opportunity. In contrast, a cap with trade policy allowed for continuous and dynamic equilibration of the value of water in irrigation diversion and carbon planting interception uses and thus provided spatially focussed incentive to avoid carbon planting in the highest water opportunity cost, high rainfall eastern and southern fringes of the MDB. Through this more efficient allocation of land use and water, the cap and trade policy allowed $2 billion dollar greater annual return in the M3 moderate climate mitigation outlook and $3.4 billion greater annual return in the L1 more aggressive mitigation outlook than in the cap without trade scenario by 2050 with an identical cap on net flow impacts. This finding was for relatively conservative assumptions about landholder response to incentive and land competition feedback on food prices. Even with more conservative assumptions, the estimated benefit value of trade provisions exceeded $1 billion annually by 2050 in many cases.

19

4.3. Policy implications The analysis illustrated large potential benefit from provisions for water trade within a cap on flow to mitigate impacts of reforestation runoff interception. However, realising the potential benefits requires significant investment in institutional infrastructure for accounting, monitoring, enforcing and facilitating efficient exchange (Dinar, 2002; Garrick et al., 2012). Illustratively, the reforms to develop the existing Murray-Darling Basin system of tradeable water diversion rights took over a decade. While no estimates of total cost involved are available, the process of producing legislation, water plans, water meters, water accounting registers, enforcing limits, and developing efficient trading platforms would have involved very significant expense (Kendall, 2013). Significant benefits of a magnitude likely to justify these costs were mostly realised more than a decade after the reforms when the trade that the reforms enabled allowed billions of dollars of benefit in decreased cost of drought (Connor and Kaczan, 2013). Incentive conditions conducive to large carbon planting response may, similarly, only arise on a long and difficult-toforesee time scale depending on Australian policy as well as the detail and degree of global accord on climate change action that emerges in coming decades. Policy to address carbon forest water interception in the MDB may not be immediately necessary given that two 2015 carbon credit auctions under the Australian Commonwealth Emissions Reduction Fund (ERF) paid $13.95 and $12.25 CO2 t1. These prices elicited substantial commitment to not clear forests and allow unassisted regrowth in more northerly and lower opportunity cost pastoral regions of Australia, but very little reforestation within the MDB region considered in this study (Bateman, 2015). The view that the prospects of realising interception cap and trade benefit may remain limited for some time to come is reinforced through the sensitivity analysis finding of very limited policy value with an assumption of very limited landholder responsiveness to carbon incentive. This finding is consistent with other research results suggesting that carbon payment uncertainty may be the greatest inhibitor of carbon planting land use change (Lagerkvist, 2005; Regan et al., 2015). The uncertainty around if and when incentive for large scale land use change could arise, leads to a challenge around choices in further water interception property rights development. As Grundy et al.

Fig. 6. 2050 increased economic returns ($AU billion) to all MDB land uses with water interception cap and trade policy (relative to cap without trade scenario) for L1 and M3 global outlooks, alternative agricultural productivity growth and land conversion hurdle assumptions and alternative land competition food price feedback assumptions.

20

J.D. Connor et al. / Environmental Science & Policy 66 (2016) 11–22

Fig. 7. Spatial distribution of land use under the three policy scenarios for both global outlooks in 2050.

J.D. Connor et al. / Environmental Science & Policy 66 (2016) 11–22

(2016) noted, landholders across large swaths of Australia’s agricultural land, especially in extensive grazing regions, have potential for much greater returns to land in a scenario of relatively high global climate action and complementary Australian policy settings encouraging mitigation through carbon sequestration. A well developed and tradeable water interception property rights system with highly spatially differentiated water entitlement requirements would allow realisation of this economic opportunity for land sector carbon sequestration while efficiently managing water opportunity costs. Less development of interception property rights can save shorter term policy implementation costs that may not be warranted if global demand for carbon sequestration doesn’t materialise. We conclude that it may well be worthwhile to further investigate costs, risks and benefits of a staged policy development strategy. A first step would involve early development of fundamental property rights systems required for trade in water interception. This could be followed by staged and incremental investment in more sophisticated monitoring, accounting, trading and enforcement systems at a pace and at locations contingent on the level of rising demand for carbon planting land use change and consequent water interception impacts. 4.4. Limitations Limits to scope of possible detail that can be modelled is an inevitable challenge with highly integrated assessments such as this one. As with all such studies we haven’t considered a number of additional benefits and costs that could arise from implementation of trade policy to address carbon forest water interception. One short coming is the somewhat simplified hydrologic representation without spatially differentiated estimates of changes in run-off from changes in interception or routing of water through river stretches, storages, and diversions. As a result our irrigation water supply impact is a global approximation for the basin as a whole and does not account for local flow routing, constraints, and water balance detail. This is likely to lead to overstatement of trade policy benefits in some locations where loss of flow may not be particularly beneficial and understatement in others areas where flow losses can produce particularly significant localised consequences. A second modelling limitation is that analysis is based on averages across years; the science shows that interception is likely to be proportionally greater in dryer years and less in wet years. Dry year impacts may be especially concerning because while water allocation to irrigation is typically adjusted downward in dry years, trees would continue to intercept water in such year and likely a greater proportion of runoff than in average years. Further, several possible location specific environmental benefits that we do not model could arise from reforestation. For example salinity, erosion control and localised increases in albedo and evapotranspiration that reduce peak flood flows and skew runoff timing to drier periods could all provide some local benefits. Addressing all of these limitations would produce both some additional costs and some additional benefits with an indeterminate net impact. However, given the overall magnitude of estimated very large total water volume consequences of interception and very large value of trade provision, we judge that the study provides a robust case for further consideration of cap and trade policy for managing potential impacts of carbon emissions abatement policy on water resources. Of course, it is also fair to conclude there are a number of areas where the modelling is limited and can be improved with additional work. 5. Conclusion A growing literature illustrates the potential for carbon pricing to significantly offset CO2 emissions through incentives for change from

21

agricultural to forest land use for carbon sequestration. A significant risk arising is the potential for adverse impacts on water supply as a result of additional water interception from such land use change. This is the first high spatial and temporal resolution, basin-scale assessment of a cap and trade approach to balance water and carbon values from carbon planting incentive policy to date, and the most comprehensive assessment of policy value sensitivity to global drivers and key parameter uncertainties. We conclude that policy to cap combined influence of irrigation water diversion and carbon planting water interception with provision for water trade has potential to significantly improve overall return to land and water relative to a cap policy without trade. The policy works by internalising what is now a water interception externality into the incentives for landholders contemplating conversion to carbon forestry. The magnitude of the potential economic efficiency gains were found to be large (billions of dollars per annum by 2050) under conditions of moderate to high carbon price and even with moderately pessimistic assumptions about landholder responsiveness to carbon price incentive. We conclude that in scenarios where markets for land based carbon gain momentum, the value in improved allocation efficiency is likely to justify the considerable cost of developing the requisite sophisticated property rights systems. While this analysis is specific to the Murray-Darling Basin, the general findings that allowing trade in a cap policy can enhance value realised from constrained water supply translates to other river basins, especially in arid to semi-arid climates where competition for water is already intense and carbon planting incentives have significant potential to further reduce flow. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j. envsci.2016.07.005. References Andreyeva, T., Long, M.W., Brownell, K.D., 2010. The impact of food prices on consumption: a systematic review of research on the price elasticity of demand for food. Am. J. Public Health 100, 216–222. Audsley, E., Trnka, M., Sabate, S., Maspons, J., Sanchez, A., Sandars, D., Balek, J., Pearn, K., 2015. Interactively modelling land profitability to estimate European agricultural and forest land use under future scenarios of climate, socioeconomics and adaptation. Clim. Change 128, 215–227. Australian Government, 2015. Carbon Farming Initiative Project Transition into the Emissions Reduction Fund. . Bateman, B., 2015. First auction underr the emisssions reduction fund. Aust. Resour. Energy J. 34, 83–86. Bathgate, A., Seddon, J., Finalyson, J., Hacker, R., 2009. Managing catchments for multiple objectives: the implications of land use change for salinity: biodiversity and economics. Anim. Prod. Sci. 49, 852–859. Benitez, P.C., Obersteiner, M., 2006. Site identification for carbon sequestration in Latin America: a grid-based economic approach. For. Policy Econ. 8, 636–651. Brown, A.E., Podger, G.M., Davidson, A.J., Dowling, T.I., Zhang, L., 2007. Predicting the impact of plantation forestry on water users at local and regional scales – an example for the Murrumbidgee River Basin, Australia. For. Ecol. Manag. 251, 82– 93. Bryan, B.A., Nolan, M., Harwood, T.D., Connor, J.D., Navarro-Garcia, J., King, D., Summers, D.M., Newth, D., Cai, Y., Grigg, N., Harman, I., Crossman, N.D., Grundy, M.J., Finnigan, J.J., Ferrier, S., Williams, K.J., Wilson, K.A., Law, E.A., HatfieldDodds, S., 2014. Supply of carbon sequestration and biodiversity services from Australia's agricultural land under global change. Global Environ. Change 28, 166–181. Bryan, B.A., Crossman, N.D., Nolan, M., LI, J., Navarro, J., Connor, J.D., 2015a. Land use efficiency: anticipating future demand for land-sector greenhouse gas emissions abatement and managing trade-offs with agriculture, water, and biodiversity. Global Change Biol. 21, 4098–4114. Bryan, B.A., Runting, R.K., Capon, T., Perring, M.P., Cunningham, S.C., Kragt, M.E., Nolan, M., Law, E.A., Renwick, A.R., Eber, S., Christian, R., Wilson, K.A., 2015b. 2015. Designer policy for carbon and biodiversity co -benefits under global change. Nat. Clim. Change . Calder, I.R., 2007. Forests and water-ensuring forest benefits outweigh water costs. For. Ecol. Manag. 251, 110–120. Chisholm, R.A., 2010. Trade-offs between ecosystem services: water and carbon in a biodiversity hotspot. Ecol. Econ. 69, 1973–1987.

22

J.D. Connor et al. / Environmental Science & Policy 66 (2016) 11–22

Chu, H.-J., Lin, Y.-P., Huang, C.-W., Hsu, C.-Y., Chen, H.-Y., 2010. Modelling the hydrologic effects of dynamic land-use change using a distributed hydrologic model and a spatial land-use allocation model. Hydrol. Processes 24, 2538– 2554. Connor, J., Kaczan, D., 2013. Principles for economically efficient and environmentally sustainable water markets: the Australian experience. In: Schwabe, K.A., Albiac, J., Connor, J., Hassan, R., Meza-Gonzale, Z.L. (Eds.), Drought in Arid and Semi-Arid Regions: A Multi-Disciplinary and Cross-Country Perspective. Springer, Dordrecht. Connor, J., Bryan, B., Nolan, M., Stock, F., Gao, L., Dunstall, S., Grahamc, P., Ernstb, A., 2015. Modelling Australian land use competition and ecosystem services with food price feedbacks at high spatial resolution. Environ. Model. Softw. 69, 141– 154. Dinar, A., 2002. Institutions, transaction costs and environmental policy. Institutional reform for water resources. Am. J. Agric. Econ. 84, 250–251. Dumortier, J., Hayes, D.J., Carriquiry, M., Dong, F.X., Du, X.D., Elobeid, A., Fabiosa, J.F., Tokgoz, S., 2011. Sensitivity of carbon emission estimates from indirect land-use change (vol. 33, p. 428, 2011). Appl. Econ. Perspect. Policy 33, 673. Dymond, J.R., Ausseil, A.-G.E., Ekanayake, J.C., Kirschbaum, M.U.F., 2012. Tradeoffs between soil, water, and carbon – a national scale analysis from New Zealand. J. Environ. Manag. 95, 124–131. Egginton, P., Beall, F., Buttle, J., 2014. Reforestation – climate change and water resource implications. For. Chron. 90, 516–524. Ewert, F., Rounsevell, M.D.A., Reginster, I., Metzger, M.J., Leemans, R., 2005. Future scenarios of European agricultural land use I. Estimating changes In crop productivity. Agric. Ecosyst. Environ. 107, 101–116. Farley, K.A., Jobbagy, E.G., Jackson, R.B., 2005. Effects of afforestation on water yield: a global synthesis with implications for policy. Global Change Biol. 11, 1565– 1576. Gao, L., Bryan, B.A., 2016. Incorporating deep uncertainty into the elementary effects method for robust global sensitivity analysis. Ecol. Modell. 321, 1–9. Gao, Y., Zhu, X., Yu, G., He, N., Wang, Q., Tian, J., 2014. Water use efficiency threshold for terrestrial ecosystem carbon sequestration in China under afforestation. Agric. Forest Meteorol. 195, 32–37. Gao, L., Bryan, B.A., Nolan, M., Connor, J.D., Song, X., Zhao, G., 2015. Robust global sensitivity analysis under deep uncertainty via scenario analysis. Environ. Model. Softw.. Garrick, D., Bark, R., Connor, J., Banerjee, O., 2012. Environmental water governance in federal rivers: opportunities and limits for subsidiarity In Australia's MurrayDarling river. Water Policy 14, 915–936. Gillet, V., Mckay, J., Keremane, G., 2014. Moving from local to state water governance to resolve a local conflict between irrigated agriculture and commercial forestry in South Australia. J. Hydrol. 519, 2456–2467. Golub, A.A., Henderson, B.B., Hertel, T.W., Gerber, P.J., Rose, S.K., Sohngen, B., 2013. Global climate policy impacts on livestock, land use, livelihoods, and food security. Proc. Natl. Acad. Sci. 110, 20894–20899. Grafton, R.Q., Horne, J., 2014. Water markets In the Murray-Darling basin. Agric. Water Manag. 145, 61–71. Grafton, R.Q., Horne, J., Wheeler, S.A., 2016. On the marketisation of water: evidence from the Murray-Darling basin, Australia. Water Resour. Manag. 30, 913–926. Grigg, N.K., Hanslow, I., Harman, I., Finnigan, J.J., 2013. Water In Ano: An Overview. Canberra. Grubb, M., 2012. Emissions trading: cap and trade finds new energy. Nature 491, 666–667. Grundy, M.J., Bryan, B.A., Nolan, M., Battaglia, M., Hatfield-Dodds, S., Connor, J.D., Keating, B.A., 2016. Scenarios for australian agricultural production and land use to 2050. Agric. Syst. 142, 70–83. Hanemann, M., 2010. Cap-and-trade: a sufficient or necessary condition for emission reduction? Oxford Rev. Econ. Policy 26, 225–252. Hatfield-Dodds, S., Brinsmead, T.S., Bryan, B.A., Graham, P., Grundy, M., Harwood, T. D., Newth, D., Schandl, H., Wonhas, A., Adams, P., Mckellar, L.B., Cai, Y., Ferrier, S., Finnigan, J., Hanslow, K., Mccallum, R., Nolan, M., Prosser, I., Smith, M.S., Baynes, T., Chiew, F., Connor, J.D., Geschke, A., Grigg, N., Harman, I., Hayward, J., Keating, B.A., King, D., Lenzen, M., Lonsdale, M., Mccrae, R., Garcia, J.N., Owen, A., Raison, J., Reedman, L.J., Smith, M.H., Summers, D.M., Whetton, P., 2015a. Csiro australian national outlook 2015: living standards, resource use, environmental performance and economic activity, 1970–2050. Technical Report, Canberra, Australia. Hatfield-Dodds, S., Schandl, H., Adams, P., Baynes, T.M., Brinsmead, T.S., Bryan, B.A., Chiew, F.H.S., Al, E., 2015b. Australia is ‘Free to choose’ economic growth and falling environmental pressures. Nature 527, 49–53. Hearne, R.R., Easter, K.W., 1997. The economic and financial gains from water markets In Chile. Agric. Econ. 15, 187–199. Hughes, N., Lawson, K., Davidson, A.J., Jackson, T., Sheng, Y., 2011. Productivity pathways: climate adjusted production frontiers for the australian broadacre cropping industry. Aust. Bureau Agric. Resour. Econ. Sci.. Jackson, R.B., Jobbagy, E.G., Avissar, R., Roy, S.B., Barrett, D.J., Cook, C.W., Farley, K.A., Le Maitre, D.C., Mccarl, B.A., Murray, B.C., 2005. Trading water for carbon with biological sequestration. Science 310, 1944–1947.

Kahil, M.T., Connor, J.D., Albiac, J., 2015. Efficient water management policies for irrigation adaptation to climate change In southern europe. Ecol. Econ. 120, 226–233. Kendall, M., 2013. Drought and its role in shaping water policy In Australia. In: Schwabe, K., Albiac, J., Connor, J.D., Hassan, R.M., González, L.M. (Eds.), Drought In Arid And Semi-Arid Regions. Springer. Kirby, M., Bark, R.H., Connor, J., Qureshi, M.E., Keyworth, S., 2014. How did irrigated agriculture in Australia's Murray? Darling basin adapt In the millennium drought?. Agric. Water Manag. 145, 154–162. Kriegler, E., O'neill, B.C., Hallegatte, S., Kram, T., Lempert, R.J., Moss, R.H., Wilbanks, T., 2012. The need for and use of socio-economic scenarios for climate change analysis: a new approach based on shared socio-economic pathways. Global Environ. Change 22, 807–822. Kruger, F., Crafford, J., Ginsburg, A., 2008. The regulation of water-use impacts of forestry in South Africa: appraisal of the development of policy and governance. Workshop On Forest Governance & Decentralization In Africa, Durban, South Africa. Lagerkvist, C.J., 2005. Agricultural policy uncertainty and farm level adjustments – the case of direct payments and incentives for farmland investment. Eur. Rev. Agric. Econ. 32, 1–23. Lane-Miller, C.C., Wheeler, S., Bjornlund, H., Connor, J., 2013. Acquiring water for the environment: lessons from natural resources management. J. Environ. Policy Plann. 15, 513–532. Lubowski, R.N., Plantinga, A.J., Stavins, R.N., 2006. Land-use change and carbon sinks: econometric estimation of the carbon sequestration supply function. J. Environ. Econ. Manag. 51, 135–152. Marinoni, O., Garcia, J.N., Marvanek, S., Prestwidge, D., Clifford, D., Laredo, L.A., 2012. Development of a system to produce maps of agricultural profit on a continental scale: an example for Australia. Agric. Syst. 105, 33–45. Newth, D., Cai, Y., Finnigan, J., Harman, I., Grigg, N., 2015. The shrinking space for climate mitigation. Australian National Outlook Working Paper, Csiro, Canberra. Nordblom, T.L., Christy, B.P., Finlayson, J.D., Roberts, A.M., Kelly, J.A., 2010. Least cost land-use changes for targeted catchment salt load and water yield impacts In south eastern Australia. Agric. Water Manag. 97, 811–823. Nordblom, T.L., Finlayson, J.D., Hume, I.H., 2012. Upstream demand for water use by new tree plantations imposes externalities on downstream irrigated agriculture and wetlands. Aust. J. Agric. Resour. Econ. 56, 455–474. NWC, 2011. The National Water Initiative—Securing Australia’s Water Future: 2011 Assessment, September 2011. Australia, C.O. (Ed.). Paltsev, S., Reilly, J.M., Jacoby, H.D., Gurgel, A.C., Metcalf, G.E., Sokolov, A.P., Holak, J. F., 2008. Assessment of us ghg cap-and-trade proposals. Climate Policy 8, 395– 420. Polglase, P., Paul, K., Hawkins, C., Siggins, A., Turner, J., Booth, T., Crawford, D., Jovanovic, T., Hobbs, T., Opie, K., Almeida, A., Carter, J., 2008. Regional Opportunities For Agroforestry Systems In Australia. Canberra, Rural Industries Research And Development Corporation. Prestemon, J.P., Wear, D.N., 2000. Linking harvest choices to timber supply. Forest Sci. 46, 377–389. Qureshi, M.E., Connor, J., Kirby, M., Mainuddin, M., 2007. Economic assessment of acquiring water for environmental flows in the Murray basin. Aust. J. Agric. Resour. Econ. 51, 283–303. Regan, C.M., Bryan, B.A., Connor, J.D., Meyer, W.S., Ostendorf, B., Zhu, Z., Bao, C., 2015. Real options analysis for land use management: methods, application, and implications for policy. J. Environ. Manag. 161, 144–152. Richards, K.R., Stokes, C., 2004. A review of forest carbon sequestration cost studies: a dozen years of research. Clim. Change 63, 1–48. Schmalensee, R., Joskow, P.L., Ellerman, A.D., Montero, J.P., Bailey, E.M., 1998. An interim evaluation of sulfur dioxide emissions trading. J. Econ. Perspect. 12, 53– 68. Schmitz, C., Biewald, A., Lotze-Campen, H., Popp, A., Dietrich, J.P., Bodirsky, B., Krause, M., Weindl, I., 2012. Trading more food: implications for land use, greenhouse gas emissions, and the food system. Global Environ. Change 22, 189–209. Schrobback, P., Adamson, D., Quiggin, J., 2011. Turning water into carbon: carbon sequestration and water flow In the Murray-Darling basin. Environ. Resour. Econ. 49, 23–45. Stavins, R.N., 1999. The costs of carbon sequestration: a revealed-preference approach. Am. Econ. Rev. 89, 994–1009. Van Dijk, A.I.J.M., Renzullo, L.J., 2011. Water resource monitoring systems and the role of satellite observations. Hydrol. Earth Syst. Sci. 15, 39–55. Van Dijk, A.I.J.M., Hairsine, P.B., Arancibia, J.P., Dowling, T.I., 2007. Reforestation water availability and stream salinity: a multi-Scale analysis in the MurrayDarling basin, Australia. For. Ecol. Manag. 251, 94–109. Young, M.D., Mccoll, J.C., 2009. Double trouble: the importance of accounting for and defining water entitlements consistent with hydrological realities. Aust. J. Agric. Resour. Econ. 53, 19–35.