Distributional impacts of water allocation policies for an agricultural watershed

Canadian Water Resources Journal Vol. 34(3): 229–244 (2009) © 2009 Canadian Water Resources AssociationRevue canadienne des ressources hydriques

Justin To1, Alfons Weersink1, and Rob de Loë2

1 Department of Food, Agriculture and Resource Economics, University of Guelph, Guelph, ON N1G 2W12 Department of Environment and Resource Studies, University of Waterloo, Waterloo, ON N2L 3G1

Submitted August 2008; accepted April 2009. Written comments on this paper will be accepted until March 2010.

Distributional Impacts of Water Allocation Policies for an

Agricultural Watershed

Justin To, Alfons Weersink, and Rob de Loë

Abstract: The study examined the short and long-run efficiency of current and alternative water allocation policies in Ontario. Information was gathered on water supply and the value of water for agricultural irrigation, domestic use and recreation/environmental purposes. A broad 10% reduction of all water takings under Ontario’s Low Water Response plan would cost the residents of the watershed approximately $1.63 million with the bulk of the cost borne by agriculture ($1.59 million). Targeting reductions to minimize costs lowers the total to less than $635,000. In the short-run, simulations showed the efficiency of current Ontario policy could be improved by specifying clear policy objectives in times of scarcity and targeting reductions to lower value uses. Simulations of alternative policies suggested maximum permitted volume fees and volumetric fees provided insufficient incentives to change water use in the short-run and only volumetric fees added incentives to invest in efficiency in the long-run. To achieve reductions in consumptive use in the short-run, volumetric fees would have to be very large and would add significant costs to users. Permit trading systems added significant incentives to allocate water to its most beneficial uses in the short-run and incentives for conservation in the long-run.

Résumé : L’étude a porté sur l’efficacité à court terme et à long terme des politiques d’affectation des ressources en eau en Ontario, tant les politiques actuelles que les politiques de rechange. Des données ont été recueillies sur l’approvisionnement en eau et sur la valeur de l’eau pour un usage domestique, à des fins d’irrigation agricole et à des fins récréatives et environnementales. Une réduction générale de 10 % de tous les prélèvements d’eau conformément au plan d’intervention en cas de baisse du niveau des eaux de l’Ontario coûterait aux résidents du bassin hydrographique environ 1,63 million de dollars, le plus gros du coût étant assumé par le secteur agricole (1,59 million de dollars). Des réductions ciblées visant à diminuer les coûts au minimum permettent de réduire le total à moins de 635 000 $. À court terme, les simulations ont révélé que l’efficacité des politiques actuelles de l’Ontario pourrait être améliorée si des objectifs clairs en matière de politiques sont précisés en période de rareté et si des réductions des usages de moindre valeur sont ciblées. Les simulations de politiques de rechange portent à croire que les frais maximums autorisés par volume et les frais volumétriques offrent des stimulants insuffisants pour modifier l’utilisation de l’eau à court terme et que seuls les frais volumétriques procurent des incitatifs pour l’investissement

230 Canadian Water Resources Journal/Revue canadienne des ressources hydriques

© 2009 Canadian Water Resources Association

dans l’efficacité à long terme. Pour parvenir à des réductions dans la consommation en eau à court terme, les frais volumétriques doivent être très élevés, ce qui se traduirait par l’ajout de coûts considérables pour les utilisateurs. Grâce aux systèmes d’échange de permis, d’importants incitatifs ont pu être ajoutés en vue de l’affectation de l’eau à ses usages les plus bénéfiques à court terme, de même que des incitatifs pour sa conservation à long terme.

Introduction

In Ontario, the demand for water continues to grow for residential, industrial, agricultural, and environmental uses. The government of Ontario predicts that by 2020, there will be significant increases in agricultural irrigation water usage from increases in food production and overall consumptive usage due to population growth (Ontario, 2003). Available water supply, however, is highly variable and, in many instances during summer periods, water supplies have not met demands placed on them putting stress on consumptive users and the environment (Kreutzwiser et al., 2004). Scarcity, driven by competing demands surpassing available supplies, requires effective management to maximize the societal benefits of water for today and into the future.

Current methods for allocating water in Ontario are regulatory based on various common and statutory laws that sometimes require withdrawal permits and sometimes not. Market mechanisms such as permit trading systems or pricing systems thus far do not exist in Ontario. Ontario’s principal mechanisms to allocate water, the Permit to Take Water (PTTW) and the Ontario Low Water Response (OLWR) programs, have left considerable ambiguity and the principles are not consistently applied. Many of the challenges, such as inadequate public participation, poorly defined priority uses, inadequate understanding of current water use and watershed supply, and confused intergovernmental jurisdiction, have been discussed in Hoffman and Mitchell (1995), Leadlay and Kreutzwiser (1999), and Kreutzwiser et al. (2004). In response to these challenges, the government of Ontario released a proposal for water policy change (Government of Ontario, 2004b) and indicated that it intended to modernize provincial water policy to encourage conservation and introduce fair pricing

of water to protect and manage water resources for the future. New regulations instituted in 2004 and 2007 (Ontario Regulations 387/04 and 450/07), and an updated policy manual (Ontario Ministry of the Environment, 2005), addressed many of these concerns. However, considerable scope for improvement remains, especially during times of shortage.

The purpose of this study was to assess the efficiency of Ontario’s current policies in allocating water and to examine the potential of alternative water allocation policies. The study examines alternative options for allocating water to maximize benefits for all users: communities, businesses, agriculture and the environment. Specifically, the study explores whether water allocation policies generate incentives for users to allocate water to its most beneficial use given fluctuations in water scarcity during the year. This is an important concern in the summer season as available water can be insufficient to meet all demands. Also important is a long-term concern: do water allocation policies generate incentives for users to continually invest to ensure they are minimizing waste and using water to maximize the benefits of its use? Long-term supply and demand for water may change due to changing preferences for environmental benefits, agricultural technology, and climatic conditions. The Big Creek Watershed in southern Ontario was used as a case study to assess the potential impacts of alternative allocation policies. A model of water users within Big Creek was constructed incorporating estimates of the various values and supply constraints within the region.

Empirical Water Allocation Model

The study analyzed the efficiency of water allocation policies based on two major criteria: 1) in the short-run, within a summer season, do water allocation policies generate incentives for users to allocate water to its most beneficial use given fluctuations in water scarcity during the year?; and 2) in the long-run, do water allocation policies generate incentives for users to continually invest in new technology and infrastructure to maximize the benefits of water in the long-run?

In the short-run, a conceptual framework was constructed to evaluate the benefits derived from consumptive and non-consumptive uses of water. Consumptive uses of water were defined as uses in which water is extracted from its source but not

To, Weersink, and de Loë 231


returned. The benefit function for consumptive user j (denoted by Bj ) is assumed to be an increasing and strictly concave function of consumption (denoted by cj ). Consumption was defined as the amount of water withdrawn less return flows (i.e., water used but not returned to its source). Non-consumptive uses were defined as uses in which value was derived from the quantity of water inherent within the source (denoted by v) and water is not removed. The benefit function of non-consumptive user i (denoted by Bi

NCU) is assumed to be increasing and strictly concave in v. The optimal allocation of water in the short-run maximizes the benefit functions of all users subject to costs and constraints on water flow. The problem can be summarized as

(1)

where N is the number of consumptive users, Pw is the per-unit price paid for water consumed by an individual, v0 is the initial flow of water and BNCU(v) is the sum of all non-consumptive user benefits. It is important to note that in the short-run analysis, the definition of consumptive water use benefit functions (Bc) and non-consumptive benefit functions (BNCU) will be based on short-run demands for water (i.e., seasonal demands) and not long-run, multi-year benefit/demand functions.

To determine the efficiency of water allocation polices in the long-run, the study focused on whether policies contributed incentives for users to invest to conserve water and improve the effectiveness of their water use. Do users have incentives such as increased profit, reduced risk or increase in net benefits that would drive the user to invest in technology or infrastructure to increase the benefits of water use? An example for agriculture would be investing in drip irrigation. Ideally, the long-run analysis would examine long-run demand and supply relationships to determine efficiency. However, this analysis would require determination of technology paths, evolution of long-run environmental and social preferences, long-run market fluctuations and other considerations. This was deemed to be beyond the capacity of this study.

Short-Run Model

The conceptual model summarized by Equation (1) involves four major components: consumptive water use benefits (Bc), non-consumptive benefits (BNCU), hydrological relationships (v, v0) and water policy (Pw or other policy interventions affecting Pw and v). In the following sections, estimates for each of these components are described for the Big Creek watershed.

Study Area

The Big Creek watershed (Figure 1), situated in southwestern Ontario, encompasses approximately 723 km2 and flows south 120 km into Lake Erie. The watershed is predominantly rural with a largely agricultural-based economy (Gamsby and Mannerow Ltd. et al., 2002). The population of Big Creek watershed was approximately 19,000 in 1996 and growth estimated at 0.1% per year. The largest population centre is the town of Delhi, with a population of 4,694 in 2002 (Waterloo Hydrogeologic et al., 2003).

Water Supply

Streamflow along Big Creek has been monitored since 1954. Over the last five decades, the monthly mean flow in the summer months of Big Creek (Walsingham gauge) ranged from 1.6 to 10.7 m3/s. In recent dry years, there has been insufficient supply to meet all water demands in July and August. This has resulted in the damming of creek flows, fish kills due to over-pumping, municipal watering bans and stakeholder conflicts over the allocation of water.

Groundwater uses in the watershed were found to consume only a fraction of groundwater recharge and are not a significant constraint when considering yearly budgets (Waterloo Hydrogeologic et al., 2003). It must be noted, however, that groundwater consumption can lower water tables in summer months, which may contribute to reduced streamflow. Determining the extent to which this occurs in specific locations is complex and depends on an array of factors, including soil type, distance to bedrock, distance to stream, hydraulic conductivity and soil density. No such estimates of groundwater/surface water interactions



Figure 1. Location of Big Creek Watershed.



have been done within the watershed, and the location-specific analysis was beyond the scope of this study. For the purposes of this study, water users were determined to be either surface water users, drawing directly from Big Creek or its tributaries, or groundwater users. Simulations attempted to assess the impact of some groundwater-surface water interactions but were not considered to be an assessment of actual Big Creek surface water-groundwater dynamics.

Consumptive Water Use

During the time the study was completed, crop irrigation was the most significant water use in Big Creek, with approximately 90% of crop area irrigated in tobacco (Table 1). Data on actual water takings for irrigation within the watershed do not exist. Surveys of agricultural water use within the watershed were attempted but the data were found to be inadequate for analysis. Many producers did not know how much water they applied and none claimed to have attempted to estimate the yield and quality improvements gained through irrigation. Further, little research has been done measuring yield response to irrigation in southern Ontario. Looking to other sources of data, one of the most prevalent methodologies for assessing yield response to irrigation in the literature is that of Doorenbos and Kassam (1979). Crop-specific yield responses using this method outline percentage yield response to percentage reductions in optimal water and therefore can be applied to different regions. Using this method, yield response (YR), the loss in yield for a given crop from less water use, was calculated as

YR = θ × YRC (2)

where θ is the percentage reduction of the crop’s total optimal water requirement and YRC is the yield response coefficient. Yield response coefficients for the major crops in the region were drawn from the literature (Allen et al., 1998) and are listed in Table 1.

Optimal water requirements for the development stages of each of the major crops were taken from the Ontario Best Management Practices Irrigation Management recommendations (Government of Ontario, 2004a). The irrigation requirements for the month of July and August were estimated by subtracting average precipitation from 1997-2004 for the months

of July (573 m3/ha) and August (489 m3/ha) from each crop’s optimal water requirements. Percentage loss in yield due to a reduction in irrigation was multiplied by regional average yields and crop prices (Ontario Ministry of Agriculture, Food and Rural Affairs, 2005) to estimate the marginal benefits of water by crop type. Costs were assumed to be fixed in the short-run so this calculation was assumed to equal the implicit loss of profit due to loss of irrigation water. The short-run impacts of a 1 m3 reduction in water use on farmer profits are summarized in Table 2.

Data showed that 42% of irrigation permit holders were estimated to be surface water users and the remainder groundwater users (Waterloo Hydrogeologic Inc., 2003). Average farm size was estimated using total area of crops and numbers of farms from the 2001 Agricultural Census. Total irrigated areas for each crop (Statistics Canada, 2001) and the average area per farm were used to calculate the total number of irrigated farms by crop type within Big Creek. It was assumed that, in every farm that irrigates, the total area of the major crop within that farm were irrigated. The resulting average size and water demand, by crop, in Big Creek are summarized in Table 3.

The results using this methodology can only be used to estimate short-run marginal benefits of water. The methodology calculates a constant marginal benefit for water, in other words, a 1 m3 reduction in water will have a constant effect on profitability. In the long-run this is a poor assumption as a farmer faced with water reductions could invest in better technologies, adapt crop types, varieties or other inputs to minimize the loss in benefits due to a loss in water. In the short-run however, little can be done by a farmer to adapt to reductions in water. Given week-to-week changes of water conditions within the growing season, farmers’ crop choices, technology and input costs are fixed and little can be done to adapt. Therefore yield losses due to reductions in water have a much more direct link to losses in profitability. However, even in the short term, adaptive measures using other inputs may increase other input costs but reduce the loss in yields, resulting in a smaller loss of revenue. While the use of this methodology can give an assessment of short-run impacts, the method will likely overestimate the true value.



Table 1. Big Creek Major Crops, Irrigated Area, Season Length and Water Requirements.

Crop Area

Irrigated

(ha)

Typical

Harvest

Date†

Growing

Season

Length

(days)

Total Water

Requirement

(m3/ha)

Irrigation

Requirement

July

(m3/ha/mo)

Irrigation

Requirement

August

(m3/ha/mo)

Tobacco 7315 Aug-Sept 115 4172.9 552.0 636.0Ginseng* (3 yr. rotation) 750 Aug-Nov 420 15240.0 552.0 636.0Sweet corn 42 July-Oct 100 3628.6 552.0 636.0Cucumbers 91 June-Oct 100 4989.3 1395.7 1479.7Potatoes 203 Sept 115 5669.6 1395.7 1479.7Pumpkins 120 July–Oct 100 4989.1 1395.7 1479.7Tomatoes 17 July–Oct 100 3492.5 805.1 889.1Peppers 51 July–Oct 100 4989.3 1395.7 1479.7Cabbage 44 June-Oct 115 5397.5 1395.7 1479.7Cauliflower 12 June-Oct 115 5397.5 1395.7 1479.7Carrots 26 July-Oct 100 4989.3 1395.7 1479.7Apples 22 Sept-Oct 115 3175.0 552.0 636.0Strawberries** (1.5 yr. rotation) 92 Mid-July 180 9797.1 1114.4 1198.5

† Typical Planting date mid-late May* Ginseng – assumed 14 growing months over three years** Strawberries – assumed six growing months

Table 2. Effect of Reducing Water Use by 1 m3 by Crop in Big Creek Watershed.

Crop Average†

Yield

(kg/ha)

Average†

Price

($/kg)

Yield Response

Coefficient

(YRC)

% Reduction in

Water Requirement

(θ)*

Reduction in Profits

($/m3)

July August

(a) (b) (c) (d) (e) (f)

Tobacco 2,556 5.00 0.9 0.000240 2.80 2.80Ginseng 3,363 38.49 1.01 0.000066 8.58 8.58Sweet corn 10,424 0.12 1.25 0.000276 0.51 0.50Cucumbers 16,140 0.34 1.01 0.000200 1.12 1.12Potatoes 21,274 0.23 1.1 0.000176 0.97 0.97Pumpkins 10,805 0.38 1.01 0.000200 0.82 0.82Tomatoes 67,565 0.13 1.05 0.000286 2.74 2.86Peppers 15,647 0.55 1.1 0.000200 1.85 2.04Cabbage 22,585 0.28 0.95 0.000185 1.25 1.15Cauliflower 17,777 0.50 1.01 0.000185 1.67 1.67Carrots 38,053 0.15 1.01 0.000200 1.15 1.15Apples 24,502 0.43 1.01 0.000315 3.54 3.54Strawberries 5,044 2.39 1.01 0.000102 1.33 1.33

† Averages are for 2000-2004* θ is equal to 1/total water requirement (column 4 of Table 1)



Table 3. Average Size and Water Demand by Crop Farms in Big Creek Watershed.

Average ha/Farm Number of

Irrigated Farms

in Big Creek

July Irrigation

Requirement per

Farm

(m3)

August Irrigation

Requirement

per Farm

(m3)

Minimum

Irrigation

Requirement

per farm for the

growing season

(m3)

Tobacco 23.59 310 13,022 15,007 28,029Ginseng 7.08 106 3,917 4,514 8,431Sweet corn 21 2 11,584 13,349 24,933Cucumbers 9.23 10 12,894 13,671 26,565Potatoes 26.75 8 37,352 39,604 76,956Pumpkins 6.39 19 8,937 9,476 18,413Tomatoes 5.06 3 4,058 4,482 8,539Peppers 5.46 9 7,619 8,078 15,697Cabbage 7.93 6 11,043 11,708 22,751Cauliflower 6.15 2 8,559 9,075 17,633Carrots 9.39 3 13,103 13,893 26,996Apples 13.48 2 7,438 8,572 16,010Strawberries 3.6 26 4,001 4,303 8,304

Municipal Use

The Town of Delhi draws water from two sources: 1) the reservoir at Lehman Dam on North Creek, a tributary to Big Creek; and 2) a municipal well located 4 km southeast of Delhi (Figure 1). In 2001, the annual water use of Delhi was 623,955 m3, of which 21% was from North Creek.

To estimate demand for water by municipal residents of Delhi, a telephone contingent valuation survey was conducted to determine an estimate of the value of water to Delhi residents for domestic use. A total of 156 households were chosen randomly from the phonebook and contacted by phone in the spring of 2005. A total of 50 households, representing 159 people in the town of Delhi, completed the survey. Water consumption in Delhi is metered and a use charge of $1.39/m3 is applied. Those surveyed were asked to reveal their willingness to pay to avoid 10% and 20% water reductions in their domestic use.

Within the literature there is considerable debate as to the utility of contingent valuation surveys to measure such benefits. Young (2005) provides an excellent summary of some of these issues. This study will not go further into this debate but accepts the risks associated with this method. The data gathered were

not intended to be exhaustive and thus are only used as estimates. Nonetheless, the analysis builds in the major parameters such that future research can improve the accuracy of the results.

The data on willingness to pay to avoid reductions were used to estimate the following inverse demand function for water

ln Pw = 5.771 - 0.0000156 c Adj R2 = 0.974. (3) (25.92) (-21.87)

where Pw is the price Delhi citizens were willing to pay for water in dollars per person, c is the amount of water cumulatively consumed by the citizens of Delhi and t-statistics for the estimated coefficients are in parentheses. Using the estimated coefficients, the own-price demand elasticity for water under current consumption and prices is -0.186, which is comparable to the residential price elasticity of -0.124 estimated for Ontario municipalities in 1991 (Renzetti, 1999).

Non-Consumptive Water Use Both natural and constructed water bodies in the watershed are used for recreation and environmental purposes. Boating, fishing, swimming and picnicking are the major recreational uses associated with various



locations in the watershed, such as the Teeterville, Lehman and Deer Creek reservoirs. Further, extensive marshes surrounding the mouth of Big Creek are nesting grounds for waterfowl. Various migrating birds rest briefly in the region in spring and fall and the area has been designated as a World Biosphere Reserve by UNESCO.

To achieve a basic assessment of the benefit of water for recreational and environmental purposes, the same 50 households who participated in the domestic use survey also participated in a recreation/environment survey. Survey participants were presented with a scenario in which local recreational areas, such as Teeterville and Lehman dams, faced 10% and 20% reduced flow levels. Participants were asked their willingness to pay to avoid the 10% and 20% reductions in water flow if they were presented with the ability to buy back water from consumptive users.

Results of the survey showed that 28.3% and 32.7% of the people were willing to pay to avoid 10% and 20% losses in streamflow, respectively. Linking the results of the survey with average seasonal flows of Big Creek, at Walsingham, the non-consumptive benefits of water (BNCU) were estimated for July and August. The estimated benefit functions were

July: BNCU, July = 54394.31 - 0.0059 vJuly Adj R2 = 0.973. (6.73) (-6.02) (4)

August:BNCU, Aug = 54394.31 - 0.0043 vAug Adj R2 = 0.973. (6.74) (-6.02) (5)

The estimated benefit functions suggested, cumulatively, Delhi citizens were willing to pay $0.0043/m3 and $0.0059/m3 for the months of July and August, respectively to maintain streamflow. In comparison, Young and Gray (1972) found recreation values of US$0.0024 to 0.004/m3; Creel and Loomis (1992) found a value of US$0.24/m3 for waterfowl hunting, fishing and wildlife viewing; and Ward (1987) found values of US$0.016 to 0.024/m3 for anglers and rafters in New Mexico. These estimates are regionally specific, dependent on the values specific to its users, their incomes and their value judgments, but they do suggest that the estimated non-consumptive use values for Delhi are comparable.

Water Policy

Current Water Policy in OntarioOntario’s two main water allocation policies during the study were the Ontario Water Resources Act (OWRA), which governs the Permit to Take Water (PTTW) Program, and the Ontario Low Water Response Plan (OLWR). The PTTW states that anyone withdrawing more than 50,000 L/day of water requires a permit. No permit is required for certain priority uses, such as domestic purposes, livestock watering when the water is not taken into storage, and firefighting. For water users requiring a permit, some application fees are required but no volumetric fees are applied. Agricultural water users are not charged an application fee.

The OLWR is designed to implement short-run strategies in the event of drought. The response plan defines three levels of low water conditions, each based on streamflow. Level I defines a potential water supply problem (monthly flow is less than 70% of the lowest average summer month flow), Level II defines a potentially serious problem (monthly flow is less than 50% of the lowest average summer month flow), and Level III defines a region that can no longer supply sufficient water to meet demands (monthly flow is less than 30% of the lowest average summer month flow). As part of the strategy, a Water Response Team is responsible for outlining targets of 10% and 20% voluntary reductions in water uses for Level I and II, respectively. Level III of the OLWR specifies only “maximum” reductions in water use and has never been enacted. This study will focus on Level I and II responses. While the response plan defines uses as essential, important and non-essential, Kreutzwiser et al. (2004) reported that the OLWR has not provided strategies to allocate water to priority uses in the event of low water conditions. Priority of water management is unclear and conflicts between agricultural, industrial, and environmental water uses have arisen, as observed in dry years in the Big Creek watershed.

Three allocation strategies under Level I and II water conditions were simulated: 1) a constant percentage reduction in water use for all users; 2) a cumulative percentage reduction in water to minimize cost to the region; and 3) the allocation of water in the region to maximize benefits of water use regardless of target reductions.



Alternative Policy OptionsIn 2004, the Government of Ontario presented a White Paper on watershed-based source protection planning and identified options for establishing fees related to water takings by commercial users. The objectives of the proposal were to increase conservation, establish minimum streamflows to protect the natural function of ecosystems, and ensure users pay their fair share to manage water resources for the future. These policy options, if implemented, could replace or work in addition to existing PTTW and OLWR policy. This study evaluated two of these options as well as a third. The options evaluated were:

1) A volume based charge applied relative to the maximum permitted volume;

2) A volumetric fee on water use adjusted for loss factors; and

3) A tradable permit system.

These alternative policy options were simulated as short-run adaptations to low water supply but also as long-run policies, introducing permanent water use pricing.

For both current and alternative water policy simulations, each was assessed for the efficiency in which water was allocated in the short-run and the incentives for efficient water use in the long-run. In the short-run analysis, using Equation (1), simulations summed the cumulative costs incurred by irrigators due to losses in water use (and/or price paid for water use), total willingness to pay to avoid losses of water for domestic use and willingness to pay to avoid losses of water for environmental/recreational use.

Simulation of Water Allocation Policies

Current Water Policy

For the purposes of brevity, only the results of the July, Level I simulations are presented below (Table 4). Results for Level II, and those in August, are available from the first author. The results of the Level II and August simulations were similar to the Level I, July results, only differing in magnitude.

Ten percent reduction in water use for all usersUnder Level I conditions the OLWR asks users within the watershed to voluntarily decrease water

consumption by 10% and does not specify location, priority uses or source of withdrawal. If all users reduced consumption by 10% the simulation showed this would impose a short-run cost of $1,630,077 for the region, assuming groundwater withdrawals do not affect streamflow. The cost impact of the 10% reduction was felt most at an aggregate level by tobacco farmers whose cumulative revenues fell by $1.13 million or $154/ha. On a per hectare basis, the largest costs were borne by ginseng farmers at $474/ha. The 10% reduction had little affect on residential users as the citizens of Delhi would have been willing to pay, cumulatively, only $1,585 to not reduce their water consumption by 10%. Delhi residents also lost $40,250 in recreational values from the lower volume of water within Big Creek.

The reductions in water withdrawal improved streamflow by 222,807 m3/mo or 7.53% above the Level I threshold. Of note, the improvement was very small compared to normal flow with streamflow remaining only one-quarter of the normal average monthly flow for July.

The assumption of no groundwater withdrawal-streamflow interaction was a challenge to the results of the simulation. Without greater knowledge of groundwater-streamflow interactions, an assessment of the true impact of groundwater withdrawals is difficult. To gain some understanding of the impact, a simulation was done estimating the effect of a 10% groundwater-surface water interaction (i.e., for every litre of groundwater withdrawal, streamflow would be reduced by 0.1 litres). This analysis would at least give an estimate as to the magnitude of cost impact if future assessments of groundwater-streamflow interactions were available. The results showed that a 10% groundwater-streamflow interaction would have very little financial impact, slightly decreasing the total cost of the reduction to $1,629,942, a difference of only $135, and increasing streamflow 254,197 m3/mo, or 8.6% above the Level I threshold.

Ten percent reduction in water use among users to minimize costsRather than expecting all users to cut back takings by the same percentage, the second simulation examined who could reduce consumption to minimize costs while meeting the goal of 10% reduction in total water takings. The simulation showed, while total water use for agriculture was reduced by approximately 10%,



Tab

le 4

. Eff

ects

of

Alt

ern

ativ

e R

esp

on

ses

to L

evel

I W

ater

Co

nd

itio

n, J

uly

.

10%

Red

uct

ion

by

All

Use

rs10

% R

edu

ctio

n in

Use

to

Min

Co

sts

Use

to

Min

imiz

e To

tal C

ost

s

Tota

l

Ben

efit

Lo

ss

($)

Tota

l

Red

uct

ion

in W

ater

(m3 /

mo

)

Rev

enu

e

Lo

ss p

er

ha

($/h

a)

Tota

l

Ben

efit

Lo

ss

($)

Tota

l

Red

uct

ion

in W

ater

(m3 /

mo

)

Rev

enu

e

Lo

ss p

er

ha

($/h

a)

Tota

l

Ben

efit

Lo

ss

($)

Tota

l

Red

uct

ion

in W

ater

(m3 /

mo

)

Rev

enu

e

Lo

ss

per

ha

($/h

a)

Toba

cco

-1,1

27,2

22-4

03,7

72-1

54.0

90

00.

000

00.

00

Gin

seng

-355

,238

-41,

399

-473

.65

00

0.00

00

0.00

Swee

t cor

n-1

,197

-2,3

33-2

8.32

-11,

967

-23,

329

-283

.16

00

0.00

Cuc

umbe

rs-1

4,12

7-1

2,62

4-1

56.1

7-6

9,72

7-6

2,30

8-1

,561

.78

00

0.00

Pota

toes

-27,

376

-28,

305

-134

.97

-273

,756

-283

,053

-1,3

49.7

90

00.

00

Pum

pkin

s-1

3,77

3-1

6,74

8-1

14.7

8-1

37,7

26-1

67,4

75-1

,147

.70

00

0.00

Tom

atoe

s-3

,817

-1,3

95-2

20.2

40

00.

000

00.

00

Pepp

ers

-13,

063

-7,0

65-2

58.0

50

00.

000

00.

00

Cab

bage

,-7

,581

-6,0

75-1

74.1

60

00.

000

00.

00

Cau

liflow

er-2

,788

-1,6

71-2

32.9

20

00.

000

00.

00

Car

rots

-4,1

49-3

,620

-159

.98

00

0.00

00

0.00

App

les-4

,258

-1,2

03-1

95.4

10

00.

000

00.

00

Stra

wber

ries

-13,

652

-10,

255

-148

.36

00

0.00

00

0.00

Tota

l Agr

icultu

re-1

,588

,241

-536

,465

-493

,176

-536

,165

00

Delh

i Con

sum

ptio

n-1

,585

-3,4

83-1

,901

-3,7

820

0

Rec

reat

iona

l/Env

ironm

enta

l-4

0,25

0-4

0,10

9-4

1,17

4

Tota

l Wat

ersh

ed C

hang

e-1

,630

,077

-539

,947

-535

,185

-539

,947

-41,

174

0

Fina

l Big

Cre

ek F

low

(m3 /s

)

1.22

71.

240

1.14

4

Tota

l Im

prov

emen

t of S

tream

flow

(m3 /m

o)

222,

807

255,

839

7,31

% In

crea

se in

Stre

amflo

w fr

om 1

.141

m3 /m

o

7.53

%8.

65%

0.25

%



reductions varied by crop type. Irrigation was allowed to continue for high-water-value crops such as tobacco, ginseng, tomatoes, peppers, cabbage, cauliflower, carrots, apples, and strawberries, but was eliminated for low-water-value crops such as sweet corn, potatoes, squash, pumpkin, and zucchini. The net result of the targeted reductions reduced total cost to the watershed from approximately $1.6 million to $535,185.

Interestingly, in this and the previous simulation, regardless of whether groundwater is assumed to have no or little affect on streamflow, the simulation reduced stream water and groundwater irrigation to achieve the 10% overall reduction target. This result is because the model was instructed to reduce water consumption by 10% wherever it could find water use, regardless of its effect on streamflow. This result is significant if one considers the overall policy objective. Assuming the objective is only to reduce water consumption by 10%, the simulation would reduce irrigation for both stream and groundwater users. However, if the policy goal is to reduce water consumption by 10% to minimize costs and to achieve maximum improvements to streamflow, reductions in irrigation would target stream water users.

Maximize Benefits to Users

Rather than requesting universal 10% reductions in water use under a Level I situation, how should water allocation change to maximize benefits to users or minimize the costs to the watershed? Results of this simulation showed no water would be allocated away from agricultural irrigators or domestic users and all costs would be borne by non-consumptive users. The total cost from the reduced streamflow to Delhi recreational users of $41,174 is significantly less than the cost of reducing water use of consumptive users. Delhi citizens were cumulatively willing to pay a maximum of $0.0043/m3 flowing within Big Creek, which is much less than the minimum value of water for agricultural irrigators ($0.50/m3 for sweet corn) and the amount paid by Delhi residents for domestic use ($1.382/m3). Even extrapolating the same individual values for recreation/environment to the entire population of the watershed (estimated 19,172 in 2005), the willingness to pay within the July, Level I simulation would remain less than $170,000 and not sufficient to reallocate water from agriculture. Only if the aggregate value of non-

consumptive use for improved streamflow was greater than $547,000 would water be allocated away from the least valuable irrigated crop.

Short-run and long-run incentives of current policyBecause current OLWR policy does not specify priorities for reductions in water use, current water policy will likely not achieve economically efficient allocations of water in the short-run. The results of the simulation showed that the manner in which the province achieves reductions can significantly lower the costs of water scarcity and maximize benefits of water use. Both in the short-run and long-run, ambiguity and uncertainty in allocation policy gives users few incentives to allocate water to its most beneficial uses, or to increase conservation and invest in new technologies to maximize water use. While uncertainty is sometimes an incentive to invest in efficiency in the long-run, the current policy, because of its ambiguity, may ask users to reduce their water use regardless of how effectively they use water. If a policy asks users to reduce regardless of whether they irrigate at peak or non-peak times, use efficient or inefficient technologies, the policy provides a disincentive to invest in better management, infrastructure and technology. Ineffective consumptive water use in turn has negative impacts on streamflow and the social and environmental benefits derived from streamflow. Therefore, specifying clear policy objectives in times of scarcity and targeting reductions to the lowest value uses has the potential to improve water allocation.

Alternative Water PoliciesThe study also examined the efficiency of several alternative water allocation policies, including the application of water use fees. The proposals presented in the Government of Ontario’s (2004) discussion paper identified options for establishing charges, citing a maximum permitted water taking fee system in British Columbia (BC) and volumetric fees like those in the United Kingdom (UK). Simply as a way to explore different pricing mechanisms, the BC and UK examples were applied directly as models of these fee systems.

The BC fee system consists of an annual rental fee based on maximum permitted water takings. Licensing is applied only to surface water users and allows licensees to take up to 49,400 m3 per year for a fixed fee of $22. A $0.61 per 1000 m3 volumetric charge is



applied beyond the first 49,400 m3. Simulating this fee system within Big Creek, (see Table 5) results showed the majority of users fell well below the basic level of 49,400 m3; thus the $22 fee was essentially an annual fixed cost and not a volumetric charge. Such a small fixed cost of $22 creates little incentive for licensees to use water efficiently if applied in the short or long-run. Given scarcity in the short-run, such a small fee gives no incentives for users to reallocate water to its most valuable uses. No consumptive users would reduce their water use given such a price and streamflow would not be improved. The small fee also does little to incent users to invest in the long-run to conserve water, improve technology and improve water use efficiency. While streamflows would not improve due to such a fee system, the province would raise $4,660 in annual rental fees.

The study also simulated a volumetric water charge similar to that used in the UK. There, a standard unit charge is applied ($0.0239/m3) and weighting factors (WF) are added according to use. Weighting factors are applied differently for surface water versus groundwater users, uses depending on time within growing seasons, and type of use such as agriculture, domestic use, or others. For example, agricultural irrigators are assigned large weighting factors as large quantities of water are removed from the system and withdrawals occur in the summer when supply concerns are highest. The net

result of the UK simulation was an applied volumetric fee of $0.0382/m3 for irrigators and $1.4498/m3 for domestic users during the summer months.

Simulation of this volumetric fee, if applied in the short-run in times of scarcity, found that the fees were not sufficient to alter irrigator behaviour as it remained below the marginal benefit of water to the least valuable crop (see Table 6). In the short-run, agricultural water use would not be reduced and streamflow would not be improved due to the fee. To truly affect water takings to improve streamflow, the relative weighting factor for irrigators, under current market prices, would have to increase to at least 14 from the current 1.6 to provide sufficient incentive. In contrast, domestic users reduced consumption by 311 m3/mo due to the water use fee.

It is more likely that the province would introduce a volumetric water fee as a permanent fee rather than a short-run supply adaptation. If used as a permanent fee, the UK model was found to raise $207,487 in revenue for the province for the month of July. While such a fee would not change agricultural water use in the short-run, in the long-run, the UK factor-adjusted volumetric fee would provide increased incentives for conservation and incentives for irrigators to improve long-run water use efficiency. At an average per farm cost of $498 and $1,353 per month for tobacco and potato farmers, respectively, this represents a more significant economic cost for water. The impact may

Table 5. Impact of Fee Linked to Maximum Permitted Volume Policy to Big Creek.

Calculated Number of

Irrigated Farms

in Big Creek

Minimum Irrigation

Requirement per farm

for the growing season

(m3)

Annual Rental

Fee Cost for

First Tier

Annual Rental

Fee Cost for

Second Tier

Tobacco 130 28,029.1 $2,860Ginseng 44 8,431.0 $968Sweet corn 1 24,932.5 $22Cucumbers 4 26,564.7 $88Potatoes 3 76,956.4 $66 $18Pumpkins 8 18,413.1 $176Tomatoes 1 8,539.0 $22Peppers 4 15,696.8 $88Cabbage 2 22,750.8 $44Cauliflower 1 17,633.4 $22Carrots 1 26,996.1 $22Apples 1 16,010.4 $22Strawberries 11 8,303.9 $242



be relatively minor; however, a 10% improvement in irrigation efficiency for the average tobacco farm would only reduce water fees paid by $107 annually, or approximately $4.54/ha/year. It is unlikely that savings of this magnitude would pay off any significant capital investment such as drip irrigation technology. Thus, careful analysis of the benefits and costs of implementing such a system is clearly warranted.

Tradable water permit systemA water permit trading system was not proposed by the Government of Ontario in its discussion paper, but is presented here as another option, using the Alberta water allocation system as a model. Ownership of water in Alberta is vested in the province, but water right permits are allocated to users through licences. Initial water allocation permits were grandfathered, allotted on the basis of historical use, and these rights were tradable, so users may sell and purchase water rights from each other. Sales may be temporary or permanent. To maintain ecosystem health and adequate allocation of water for recreation, minimum in-stream flow constraints were set within watersheds.

In the simulation, it was assumed that current users within Big Creek were also grandfathered permits

for the normal quantity of water they currently use. Current average flows for Big Creek (4.85 m3/s and 3.52 m3/s, at Walsingham for the months of July and August, respectively) were assumed to be sufficient to meet in-stream flow constraints and all users needs in an average season. As a result, no trades would occur and water use would not change. The impacts of the trading system only become significant when the system is shocked with a change in water supply or demand.

To understand the impacts of a water trading system, the study simulated a low water supply shock. The study simulated a policy in which the Ontario government set a minimum ecological, in-stream flow of 3.1 m3/s for Big Creek, and in the growing season, streamflow fell to 2.7 m3/s. The simulation instructed the maximization model to return the flow of Big Creek to 3.1 m3/s but minimize costs to users. Significant reductions by surface water irrigators were required, assuming groundwater takings had no impact on surface water flows. Irrigation was reduced to zero for all surface water takers except apple and ginseng producers. Tobacco producers were also required to reduce their water use. Groundwater irrigators were not asked to reduce their water use as they had no

Table 6. Impact of Volumetric Fees for Water Takings with Weighting Factors in Big Creek, July.

Total Volumetric Charges

($)

Total Reduction in Water Use

(m3/mo)

Tobacco -$154,378 0Ginseng -$15,828 0Sweet corn -$892 0Cucumbers -$4,827 0Potatoes -$10,822 0Pumpkins -$6,403 0Tomatoes -$534 0Peppers -$2,701 0Cabbage -$2,323 0Cauliflower -$639 0Carrots -$1,384 0Apples -$460 0Strawberries -$3,921 0Delhi Consumption -$2,386 -311Recreational/Environmental -$0.96Total Watershed Change -$207,499 -311Improvement of Big Creek Flow (m3/s) 0.0000034Total Provincial Revenue ($/mo) $207,487



affect on streamflow. Overall, the volume of water used by agricultural irrigation was reduced by 18%. Delhi domestic users did not change consumption levels, but switched all takings to groundwater, reducing surface water takings to ameliorate streamflow (see Table 7).

Despite losing the ability to irrigate, it was in the best interest of surface water users to engage in water trading to minimize the costs to their operations. In the second iteration of the simulation, the model was instructed to maximize benefits of users given the new constraints on water consumption to maintain streamflow. Assuming zero transaction costs and Delhi domestic users were not permitted to trade water rights, groundwater users of low value crops sold water to higher value surface-water irrigators. All water purchases were made by tobacco growers and the maximum market clearing price for the short-run supply shock was approximately $2.80/m3. The market clearing price is the price required to exhaust all mutually beneficial trades, where the marginal benefit of buyers and marginal cost of sellers is equated. Actual prices for trades would occur somewhere between each water permit holder’s minimum selling price (equal to the user’s loss in revenue per cubic metre) and the market clearing price. Overall the trading of water rights in response to drought allocated water to its highest valued users and reduced the overall loss in the watershed by $870,000.

In response to the short-run supply shock, the trading of water permits allowed the re-allocation of water to its most beneficial uses, water users had incentive to conserve to take advantage of trades and streamflow was ameliorated. In addition, the trading model also provided long-run incentives for consumptive users to continually improve their water use efficiency. In doing so, users can either sell excess rights or reduce their purchase needs for more permits due to changes in growth, technology or supply shocks. In both the short and long-run, the trading system was very effective in providing incentives for users to allocate water to its most beneficial uses, invest in conservation and protect minimum in-stream flows. Implementing such a system would require a complete overhaul of water allocation in the province and should only be considered following careful analysis of benefits and costs.

Conclusions

The objective of this study was to assess the short-run and long-run efficiency of Ontario’s current policies in allocating water and examine the potential of alternative water allocation policies. The Big Creek watershed in the region of Norfolk County was used as a case study. The impacts of alternative criteria for the Ontario Low Water Response plan were simulated. A broad 10% reduction in all water takings was found to cost the residents of Big Creek approximately $1.63 million, with the bulk of the cost borne by agriculture. In contrast, a reduction policy targeted to minimize costs to the region achieved the same reduction in total water takings but with total costs of less than $635,000.

Because the Ontario Low Water Response policy does not specify priority uses, current policy in Ontario provides little incentive for users to allocate water to its most beneficial uses in the short-run and little incentive to be efficient in their water use in the long-run. With ambiguous and uncertain policy, users could be asked to reduce water use regardless of the efficiency or value of their water use, giving users a disincentive to invest in better management, infrastructure and technology. This has negative impacts on streamflow and the social and environmental benefits derived from streamflow. Policy could be improved by specifying clear objectives in times of scarcity and targeting reductions to the lowest value uses.

Simulations of alternative policies suggest a maximum permitted volume fee, such as that used in British Columbia, does not provide incentives for users to allocate water to its most beneficial uses in the short-run, nor incentives to improve water efficiency in the long-run. Targeted volumetric fees such as those used in the United Kingdom were found to contribute small incentives to improve long-run efficiency. Volumetric fees, however, contributed little to improving allocation to the most beneficial uses in the short-run. To achieve reductions in consumptive use in the short-run, volumetric fees would have to be very large and would add significant costs to users.

A third alternative policy, a tradable permit model, was found to add significant benefits for efficiency in the short-run and long-run. Adaptations in water uses were made in response to low supply conditions, allocating water to its most beneficial uses. The trading model added incentives for conservation



Table 7. Water Allocation to Minimize Cost in Big Creek Watershed of Meeting Streamflow constraint of 3.1 m3/s

without and with Water Permit Market, July.

Without Water Market With Water Permit Market

Water

Source

Change in

Benefit

Total

Reduction in

Water

(m3/mo)

Water

traded

(m3/mo)

Cost or

Revenue from

Sale

Change in

Crop

Revenue

Net Gain

from

Trade

Overall

Change in

Benefits

TobaccoGround $0 0 - - - - $0Surface -$1,714,972 -614,305 522,526 -$1,458,750 $1,458,750 $0 -$1,714,972

GinsengGround $0 0 - - - - $0Surface $0 0 - - - - $0

Sweet cornGround $0 0 -13,531 $37,774 -$6,941 $30,833 $30,833Surface -$5,026 -9,798 - - - - -$5,026

CucumbersGround $0 0 -73,221 $204,414 -$81,939 $122,475 $122,475Surface -$59,335 -53,022 - - - - -$59,335

PotatoesGround $0 0 -164,171 $458,320 -$158,778 $299,542 $299,542Surface -$114,977 -118,882 - - - - -$114,977

PumpkinsGround $0 0 -97,136 $271,176 -$79,881 $191,295 $191,295Surface -$57,845 -70,340 - - - - -$57,845

TomatoesGround $0 0 -8,094 $22,595 -$22,141 $455 $455Surface -$16,033 -5,861 - - - - -$16,033

PeppersGround $0 0 -40,976 $114,394 -$75,766 $38,628 $38,628Surface -$54,865 -29,672 - - - - -$54,865

CabbageGround $0 0 -35,235 $98,366 -$43,969 $54,397 $54,397Surface -$31,840 -25,515 - - - - -$31,840

CauliflowerGround $0 0 -9,690 $27,052 -$16,172 $10,880 $10,880Surface -$11,711 -7,017 - - - - -$11,711

CarrotsGround $0 0 -20,994 $58,610 -$24,066 $34,544 $34,544Surface -$17,427 -15,203 - - - - -$17,427

ApplesGround $0 0 - - - - $0Surface $0 0 - - - - $0

StrawberriesGround $0 0 -59,479 $166,050 -$79,183 $86,867 $86,867Surface -$57,339 -43,071 - - - - -$57,339

Delhi Consumption

Ground $0 7,314 $0Surface -7,314 $0

Final Big Creek Flow –Walsingham (m3/s) 3.1 3.1Recreational/Environmental change in benefits -$19,598 -$19,598Total Watershed Change -$2,160,969 -$1,291,054

and investing in efficient water use, as users are able to sell excess water rights from gains in efficiency. If permits are grandfathered, a tradable permit system could achieve strong incentives for efficient water use while minimizing the cost to users. Combined with

the setting of minimum ecological, in-stream flow constraints, the environment can be also be protected. However, implementation of this model would require a complete overhaul of Ontario’s water allocation laws. Thus, before being considered as a policy option a



detailed study would be required to evaluate its benefits and costs.

References

Allen, R. G., L. S. Pereira, D. Raes, and M. Smith. 1998. Crop evapotranspiration – Guidelines for computing crop water requirements – FAO Irrigation and drainage paper 56. FAO – Food and Agriculture Organization of the United Nations, Rome.

Creel, M. and J. Loomis. 1992. Recreation value of water to wetlands in the San Joaquin Valley: Linked multinomial logit and count trip frequency models. Water Resources Research 28 (10): 2597-2606.

Doorenbos, J. and A. Kassam. 1979. Yield response to water. FAO Irrigation and Drainage Paper 33. Rome.

Gamsby and Mannerow Ltd. Applegate Groundwater Consultants, Waterloo Hydrogeologic Inc., Harold Schroeter Associates Ltd. 2002. Big Creek basin water budget study.

Government of Ontario. 2004a. Best management practices: Irrigation management.

Government of Ontario. 2004b. White Paper on watershed-based source protection planning, February 2004.

Hoffman, N. and B. Mitchell. 1995. Evolving toward participatory water management: The permit to take water program and commercial bottling in Ontario. Canadian Water Resources Journal 20 (2): 91-100.

Kreutzwiser, R. D., R. C. de Loë, J. Durley, and C. Priddle. 2004. Water allocation and the permit to take water program in Ontario: Challenges and opportunities, Canadian Water Resources Journal 29 (2): 135-146.

Leadlay, H. J. and R. D. Kreutzwiser. 1999. Rural water supply allocation in Ontario: An evaluation of current policy and practice. Canadian Water Resources Journal 24 (1): 1-14.

Ontario Ministry of Agriculture, Food and Rural Affairs. 2005. Horticulture Statistics, http://www.omafra.gov.on.ca/english/stats/welcome.html.

Ontario Ministry of Natural Resources, Ontario Ministry of the Environment, Ontario Ministry of Agriculture and Food, Ontario Ministry of Municipal Affairs and Housing, Ontario Ministry of Enterprise, Opportunity and Innovation, Association of Municipalities of Ontario, and Conservation Ontario. 2003. Ontario low water response: revised, Ontario: Province of Ontario.

Ontario Ministry of the Environment. 2005. Permit to take water (PTTW) Manual, Toronto, Ontario: Queen’s Printer for Ontario.

Renzetti, S. 1999. Municipal water supply and sewage treatment costs, prices, and distortions. Canadian Journal of Economics 32 (3): 688-704.

Statistics Canada. 2001. Census of Agriculture, http://www.statcan.ca/english/agcensus2001/index.htm.

Ward, F. A. 1987. Economics of water allocation to instream uses in a fully appropriated river basin: Evidence from a New Mexico wild river. Water Resources Research 23 (3): 281-392.

Waterloo Hydrogeologic Inc., Applegate Groundwater Consultants, Gamsby and Mannerow Ltd., K. Bruce MacDonald Consulting, MacViro Consultants Inc., Tunnock Consulting Inc. 2003. Norfolk Municipal Groundwater Study, May 2003.

Young, R. 2005. Determining the economic value of water: Concepts and methods. Washington, D.C., Resources for the Future.

Young, R. A. and S. L. Gray. 1972. Economic value of water: Concepts and empirical estimates, final report to the National Water Quality Commission., Contract NWC -70-280, Colorado State University, Fort Collins, March.

Distributional impacts of water allocation policies for an agricultural watershed

Documents