The Okanagan Fish-Water Management (OKFWM) Tool: Balancing Water Objectives in Real Time

This article was downloaded by: [Kim Hyatt]On: 04 March 2015, At: 09:34Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,37-41 Mortimer Street, London W1T 3JH, UK

Click for updates

Canadian Water Resources Journal / Revue canadiennedes ressources hydriquesPublication details, including instructions for authors and subscription information:http://www.tandfonline.com/loi/tcwr20

A decision support system for improving “fishfriendly” flow compliance in the regulated OkanaganLake and River System of British ColumbiaKim D. Hyatta, Clint A. D. Alexanderb & Margot M. Stockwellaa Science Branch, Fisheries and Oceans Canada, Nanaimo, British Columbia, Canadab ESSA Technologies Ltd, Vancouver, British Columbia, CanadaPublished online: 27 Feb 2015.

To cite this article: Kim D. Hyatt, Clint A. D. Alexander & Margot M. Stockwell (2015) A decision support system forimproving “fish friendly” flow compliance in the regulated Okanagan Lake and River System of British Columbia, CanadianWater Resources Journal / Revue canadienne des ressources hydriques, 40:1, 87-110, DOI: 10.1080/07011784.2014.985510

To link to this article: http://dx.doi.org/10.1080/07011784.2014.985510

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) containedin the publications on our platform. However, Taylor & Francis, our agents, and our licensors make norepresentations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of theContent. Any opinions and views expressed in this publication are the opinions and views of the authors, andare not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon andshould be independently verified with primary sources of information. Taylor and Francis shall not be liable forany losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoeveror howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use ofthe Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematicreproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in anyform to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

http://crossmark.crossref.org/dialog/?doi=10.1080/07011784.2014.985510&domain=pdf&date_stamp=2015-02-27

http://www.tandfonline.com/loi/tcwr20

http://www.tandfonline.com/action/showCitFormats?doi=10.1080/07011784.2014.985510

http://dx.doi.org/10.1080/07011784.2014.985510

http://www.tandfonline.com/page/terms-and-conditions

http://www.tandfonline.com/page/terms-and-conditions

A decision support system for improving “fish friendly” flow compliance in the regulatedOkanagan Lake and River System of British Columbia

Kim D. Hyatta,†*, Clint A. D. Alexanderb,†* and Margot M. Stockwella

aScience Branch, Fisheries and Oceans Canada, Nanaimo, British Columbia, Canada; bESSA Technologies Ltd, Vancouver, BritishColumbia, Canada

(Received 31 August 2013; accepted 20 September 2014)

The Okanagan Valley of British Columbia (BC) has one of the lowest per-capita water supplies in Canada. Large fluctu-ations in water supplies routinely induce seasonal extremes of flood and drought conditions that challenge the ability ofdecision makers to operate complex water management infrastructure (dams, dikes, irrigation networks, flood controlchannels) to satisfy competing objectives to meet human-system versus natural-system needs (e.g. protect property, irri-gate land, protect aquatic biota). An audit of water management performance from 1982 to 1997 indicated frequent non-compliance of water regulation decisions with “fish friendly,” lake level and river discharge ranges specified by the 1982Canada–BC Okanagan Basin Implementation Agreement (OBIA). Development and deployment of an environmentaldecision support system (EDSS) to provide real-time fish and water management tools (FWMT) to decision makersoffered a potential means to improve the balance of water management decisions affecting both human and naturalsystems. The resultant FWMT-EDSS described here includes: a coupled set of four biophysical models of critical rela-tionships among climate, fish and water that interact with a fifth water management rules model used to predict potentialconsequences of decisions for fish and other water users; a network of stations providing nearly instantaneous observa-tions of lake elevation, river-discharge, precipitation and snowpack; a Structured Query Language (SQL) server database;an internet-accessible, graphical user interface; and a set of end users representing decision makers from governmentagencies, industry and local communities. FWMT provides a risk assessment framework to integrate biophysical pro-cesses, deal with multiple species and geographic locations, anticipate socioeconomic outcomes of water managementdecisions and increase cooperation among water users to improve fish and water management. Comparisons of observa-tions from pre-FWMT “control” versus FWMT deployment years (n = 20 and 11, respectively) indicate significantimprovements at both daily (p < 0.001) and annual (p < 0.05) time scales in compliance of water management decisionswith OBIA guidelines to protect salmon during critical egg-to-fry emergence stages. These FWMT-enabled improve-ments were achieved without any increased damage to water system infrastructure, riparian property or agriculturalproduction from flood or drought conditions.

L’approvisionnement en eau dans la vallée de l’Okanagan de la Colombie-Britannique est l’une des plus faibles par habi-tant au Canada. Les importantes variations dans l’approvisionnement en eau entraînent des conditions saisonnièresextrêmes propices aux inondations et aux sécheresses qui éprouvent la capacité des décideurs à exploiter l’infrastructurecomplexe de gestion de l’eau (barrages, digues, réseaux d’irrigation, canaux de contrôle des crues) de façon à gérer desobjectifs concurrents pour répondre à la fois aux besoins du système humain et du système naturel (p. ex. protection despropriétés, irrigation des terres, protection du biote aquatique). Une vérification du rendement en matière de gestion del’eau effectuée de 1982 à 1997 a révélé que les décisions relatives à la régulation des eaux contrevenaient souvent auxniveaux du lac et à la gamme de déversements réputés sans danger pour le poisson énoncés dans l’Accord de mise enœuvre du bassin de l’Okanagan (OBIA) conclu en 1982 entre le Canada et la Colombie-Britannique. Le développementet la mise en œuvre d’un système d’aide aux décisions environnementales visant à fournir aux décideurs, en temps réel,des outils de gestion du poisson et de l’eau constituent peut-être un moyen d’améliorer l’équilibre des décisions enmatière de gestion de l’eau ayant une incidence à la fois sur les systèmes humain et naturel. Les outils de gestion dupoisson et de l’eau qui découlent du système d’aide aux décisions environnementales décrit ici comprennent : un ensem-ble de quatre modèles biophysiques des relations essentielles entre le climat, le poisson et l’eau qui interagissent avec uncinquième modèle de règlements en matière de gestion de l’eau servant à prévoir les conséquences potentielles desdécisions sur les pêcheurs et les autres utilisateurs de l’eau; un réseau de stations fournissant des observations presqueinstantanées des niveaux du lac, du début fluvial, des précipitations et de l’accumulation de neige; une base de donnéeshébergée sur serveur Structured Query Language (SQL; traduisez Langage de requêtes structuré); une interface utilisateurgraphique permettant d’accéder à Internet; et un ensemble d’utilisateurs finaux représentant les décideurs des organismesgouvernementaux, de l’industrie et des collectivités locales. Le système d’outils de gestion du poisson et de l’eau offreun cadre d’évaluation des risques visant à intégrer les processus biophysiques, à gérer le grand nombre d’espèces etd’emplacements géographiques, à anticiper les résultats socioéconomiques des décisions liées à la gestion de l’eau et àstimuler la collaboration entre les utilisateurs de l’eau afin d’améliorer la gestion du poisson et de l’eau. Lorsque

*Corresponding author. Email: [email protected]†Member of CWRA

© 2015 Her Majesty in Right of Canada

Canadian Water Resources Journal / Revue canadienne des ressources hydriques, 2015Vol. 40, No. 1, 87–110, http://dx.doi.org/10.1080/07011784.2014.985510

Dow

nloa

ded

by [

Kim

Hya

tt] a

t 09:

34 0

4 M

arch

201

5

mailto:[email protected]

http://dx.doi.org/10.1080/07011784.2014.985510

comparées, les observations recueillies avant la mise en œuvre des outils de gestion du poisson et de l’eau et cellesobtenues lors de leur déploiement (n = 20 et 11, respectivement) révèlent que l’utilisation de ces outils a permis d’améli-orer de façon significative aux échelles temporelles quotidienne (P < 0,001) et annuelle (P < 0,05) la conformité desdécisions en matière de gestion de l’eau avec les lignes directrices de l’OBIA sur la protection du saumon pendant lespériodes critiques d’incubation des œufs et de l’émergence des alevins. Ces outils de gestion du poisson et de l’eau ontpermis des améliorations sans augmenter les dommages aux infrastructures de traitement des eaux, aux propriétésriveraines ou à la production agricole causés par les conditions d’inondation ou de sécheresse.

Introduction

The Okanagan Valley in the southern interior of BritishColumbia (BC) (Figure 1) exhibits a continental climatecharacterized by long, hot summers and cold, relativelydry winters. Because it exists in the rain shadow of anextensive coastal mountain range, it is part of an aridregion that receives an average annual precipitation of

about 300–400 mm/year along the main valley bottom,and 800–1000 mm/year at the highest elevations of itssurrounding plateaus (Summit Environmental Consul-tants Ltd. 2010). Winter storage and spring melt ofsnowpack at higher elevations dominate Okanagan hy-drographs. A spring freshet that occurs from April toJune accounts for as much as 90% of annual inflows to

Figure 1. Location of the Okanagan Valley straddling the Canada–US border between British Columbia and Washington State(inset); locations of the Okanagan watershed boundary, valley-bottom lakes and key environmental monitoring sites are also identified(main figure). Kelowna, with 106,000 people, is the largest urban centre with extensive riparian infrastructure, followed by Penticton(31,909), Okanagan Falls (6005), Osoyoos (4845) and Oliver (4370).

88 K.D. Hyatt et al.

Dow

nloa

ded

by [

Kim

Hya

tt] a

t 09:

34 0

4 M

arch

201

5

the Okanagan Lake and River System (OLRS) in thevalley bottom (Dobson 2004). By July, the freshetdeclines and inflows to the OLRS remain low for theremainder of the summer, fall and winter interval. Largefluctuations in water supplies from rain and snowmeltinduce not only seasonal but also multi-year extremes ofdrought and flood, such that net annual inflows toOkanagan Lake during the 1921–2002 interval have dif-fered by more than an order of magnitude, from 78 mil-lion to 1.4 billion cubic metres of water (Alexanderet al. 2008).

As of 2009, the Okanagan Basin was home to atleast 351,000 people distributed among urban centres,recreational developments and farmland from Vernon inthe north, to Osoyoos in the south near the Canada–USborder (Figure 1). Although the modern-day economy ofthe Okanagan was initially driven by its development asan important agricultural area (e.g. for production of treefruits and wine) in BC, the current economy is also dri-ven by its status as a coveted Canadian destination forrecreational and retirement activities (Senese 2010). Allof these activities require an abundant supply of high-quality water to flourish, and yet the Okanagan isreported to have the lowest per-capita water supply ofany basin in Canada (Summit Environmental ConsultantsLtd. and Polar Geosciences Ltd. 2009). Moreover, theomnipresent threats of not only seasonal to annualdrought but also flood conditions have necessitateddevelopment of a complex infrastructure of dams, dikesand water distribution systems, the management of whichhas been the subject of controversy and ongoing evolu-tion since the early 1900s (Symonds 2000). This contin-ues today, wherein the Okanagan represents a microcosmfor global trends in which water managers are faced withthe daunting challenge of managing seasonal to annualvariability in water supplies to meet not only human sys-tem needs (e.g. minimize impacts of floods or droughton human system assets), but also increasingly to addressnatural system needs (e.g. maintain ecosystem structure,biodiversity and function) where the two may conflict(Richter et al. 2003; Bradford 2008).

Recognition of the potential for conflict, given multi-ple objectives, emerged early on in the history of watermanagement in the Okanagan, with the development ofriver channelization, low head dams and water distribu-tion systems of ever-increasing complexity to respond tonavigation needs, agricultural demands and drought andflood risk, plus the maintenance of critical habitat forhighly valued fish species in lake and riverine ecosys-tems. The authority for the management of water to meethuman system needs (e.g. water extraction, mitigation offlood or drought risk) lies with the government of BritishColumbia, while the authority for dealing with issuespertaining to navigable waters, Canada–US boundarywaters, and protection of natural-system needs (e.g.

maintain habitat supporting fisheries and threatened orendangered aquatic biota) lies with the government ofCanada. Consequently, there is a long history (reviewedby Symonds 2000) of federal–provincial engagement onwater management issues in the Okanagan, involving noless than eight federal–provincial Boards of Inquiry ormajor studies (in 1908–1909, 1914–1915, 1928, 1935,1943, 1974, 2005 and 2010) and the resultant formalmemoranda of agreements between Canada and BritishColumbia (1950, 1960, 1969, 1976).

The 1950 Canada–BC agreement for the constructionof major water regulation works on the OLRS in particu-lar provoked expressions of concern from US state andfederal agencies regarding the potential impacts of theseworks on future production of sockeye salmon(Oncorhynchus nerka) returning through US portions ofthe Columbia River to spawn in the Okanagan Rivernear Oliver, BC (Figure 1). Given its relevance under theterms of the Canada–US Boundary Waters Treaty, thesubject was referred to the International Joint Commis-sion, which in 1952 provided several recommendationsfor the construction and operation of the water regulationproject to avoid damaging the salmon run (Hourstonet al. 1954). Canadian and US agencies were alsoadvised to conduct additional studies regarding the iden-tification of prescriptive actions to achieve this end. Sub-sequently, agency personnel from Canada and the USexecuted extensive field studies and a synthesis of sock-eye salmon environmental requirements at several life-history stages including: adult migration, spawning, egg-and-fry incubation, fry emergence and smolt migration.Results from this work (Hourston et al. 1954) were usedto identify the timing and magnitude of Okanagan Riverflows likely to disrupt key life-history events as a basisfor predicting the impact of the proposed water regula-tion project on Okanagan sockeye.

Construction of the new Okanagan water regulationsystem was completed in 1958, and in 1969 the Canada–BC Okanagan Basin Agreement (OBA) was signed withthe stated purpose of developing a comprehensive watermanagement framework. The purpose of the frameworkplan was to “achieve a desirable balance” among threemajor goals: (1) development of water and relatedresources as required to ensure a viable economic basein the Okanagan; (2) maintenance and enhancement ofthe quality of the natural environment through manage-ment and protection of water and related resource sys-tems such as fisheries, wildlife and recreational areas;and (3) achievement of social progress in the Okanaganby creating a more equitable distribution of income,employment and opportunity between regions within thebasin (Anonymous 1974). Of particular relevance here, areport associated with OBA Task 162 used the results ofHourston et al. (1954), supplemented with additionalfield survey data and theoretical modeling (Anonymous

Canadian Water Resources Journal / Revue canadienne des ressources hydriques 89

Dow

nloa

ded

by [

Kim

Hya

tt] a

t 09:

34 0

4 M

arch

201

5

1973), to specify preferred ranges of regulated dischargeat Oliver to support the various life-history stages ofsockeye salmon, from spawning adults to emergent andmigrating fry. Specific targets (Table 1) were formallyadopted as part of the 1976 Okanagan Basin Implemen-tation Agreement (OBIA; Anonymous 1982), whichspecified that “water requirements for sockeye salmon inthe Okanagan River will be met in all years except con-secutive drought years.”

In September of 1982, management responsibilitiesfor the Okanagan Flood Control System were transferredto the regional office of the BC Ministry of Environmentin Penticton. Thus, the challenge of achieving a balanceof water regulation benefits envisaged by the OBIA,including maintenance of “fish friendly” lake levels andriver flows, fell to local water managers. An audit ofwater management performance between 1982 and 1997identified that over this 16-year interval, river dischargeexceeded recommended flow ranges for sockeye salmonin 13 years of adult migration, 7 years of spawning and7 years of egg incubation and fry migration (Bull 1999).Discussion of these results with local water managerssuggested that difficulties in maintaining compliancewith fish-friendly flow recommendations of the OBIAwere related to the complexity of balancing fisheries,flood control and water allocation benefits through theyear, given large uncertainties in: (1) forecasts of annualand seasonal water supplies; (2) the exact timing of sal-mon life-history events (spawning, egg incubation, etc.)that control their vulnerability in a particular year tolosses from flood-and-scour or drought-and-desiccationprocesses; (3) the magnitude of fish losses likely to be

caused by deviations from recommended lake level orriver flow ranges (e.g. during flood or drought condi-tions; Summit Environmental Consultants Ltd. 2002);and (4) simultaneous risks of “significant” propertylosses associated with seasonal maintenance of fish-friendly lake elevation and river discharge levels giveneither flood or drought events. Consequently, althoughthe intent of the OBIA clearly included maintenance offisheries and natural ecosystem values, the means to doso, even in this relatively data-rich system, eluded front-line fisheries and water managers.

Emergence of environmental decision support systems(EDSSs)

Interactive computer technologies and decision supportsystems (DSSs) for studying complex water resourceproblems began to appear in the mid-1970s and havebeen discussed in the water resource literature since themid-1980s (Watkins and McKinney 1995; Metcalfe et al.2005; Ahmad and Simonovic 2006; Turon et al. 2007).According to Simonovic (quoted on page 392 in Ahmadand Simonovic 2006), a decision support system

allows decision-makers to combine personal judgmentwith computer output, in a user-machine interface, toproduce meaningful information for support in a deci-sion-making process. Such systems are capable of assist-ing in [the] solution of all problems (structured, semi-structured, and unstructured) using all information avail-able on request. They use quantitative models and data-base elements for problem solving. They are an integralpart of the decision-makers’ approach to problem identi-fication and solution.

Table 1. Preferred fishery flows, for the Okanagan River at Oliver, BC, and recommended within the Okanagan Basin Water Agree-ment of 1982.

Sockeye Salmonlife historystage Dates

Preferredrange

Additional information Rationale(m3/sec)

Adult migration 1 August–15 September 8.5–12.7 Engineered concrete weirs in theflood channel create velocitybarriers to fish migration at eithervery low or very high discharge

Flow range required to permitready passage at constrictions inthe engineered, flood controlchannel

Spawning 16 September–31 October 9.9–15.6 Critical spawning habitat isconcentrated in a few km of un-channelized river

Discharge range required to ensureaccess to 90–98% of high qualityspawning habitat

Incubation 1 November–15 February 5.0–28.3 Incubation flows > or = 50% offlows during the spawninginterval

Discharge range within whichdesiccation (< 5 m3/s) or flood-and-scour losses (> 28.3 m3/s) ofeggs and alevins may be avoided

Fry migration 16 February–30 April 5.0–28.3 After 1 February, flood controltakes precedence and 28.3 m3/smay be exceeded

Flood control is given higherpriority than fish protection in theface of a clear risk of flood levelscapable of inducing major damageto riparian “property”


Dow

nloa

ded

by [

Kim

Hya

tt] a

t 09:

34 0

4 M

arch

201

5

The general structure of an environmental decision sup-port system (EDSS) involves: (1) an explicit design tosolve a range of structured to unstructured problems; (2)a powerful and user-friendly computer interface; (3) ana-lytical algorithms and models that readily incorporaterapidly changing data; (4) an ability to quickly explore achanging “solution-space” involving management forobjectives that sometimes conflict; (5) elements to sup-port a diversity of decision process styles: (6) a recursiveproblem-solving process (Segrera et al. 2003).

Development and application of DSSs are commonin the field of engineering, while environmental applica-tions have focused principally on wastewater treatmentand water-quality maintenance issues (references in Tu-ron et al. 2007). However, the development of EDSSsfor routine use to facilitate management for biodiversityvalues or ecosystem integrity is relatively rare (e.g.Zhang et al. 2014). Thus, EDSSs offer an under-utilizedapproach to tackling problems of complex and uncertainsystems like achieving a balance of objectives for watermanagement in the OLRS. This paper describes an inte-grated EDSS that covers all phases of the fish and watermanagement process in the OLRS. The EDSS is calledthe Fish and Water Management Tool DSS (hereafterFWMT). It is built on the Microsoft®.NET Framework,utilizing ASP.NET for the web browser user interface,VB.NET for application (coupled sub-models) logic andMicrosoft® Excel for downloadable output reports. Alldata are housed in a single Structured Query Language(SQL) Server 2000 relational database. A .NET Web ser-vice developed and managed by Environment Canadasupplies the real-time data to the FWMT database forsubsequent processing into daily average, minimum andmaximum values within the FWMT application. Thispaper describes the process used for FWMT develop-ment, identifies its structure, summarizes its functionalproperties and provides results of the extent to whichdeployment of FWMT has influenced compliance withOBIA guidelines for maintenance of “fish-friendly” flowsrelative to a pre-deployment, control interval.

Fish and water management issues and the processfor FWMT development

The FWMT was designed as a tool to support front-lineresource managers to improve their ability to balancemultiple objectives specified by the Canada–BC OBA.To this end, both published and unpublished reports onOLRS water management issues were reviewed, severalexpert group workshops were conducted in the Okana-gan and individual interviews with fisheries and watermanagers were completed between 1999 and 2002 toidentify: (1) the most common fish and water manage-ment problems in the OLRS, (2) the specific causesbehind these, (3) monitoring programs and indicators to

anticipate the occurrence of such problems and (4) therange of preventive or corrective actions available toresolve water regulation issues.

At the most general level, the maintenance of fish-friendly flows in the OLRS depends critically on manag-ing variations in the timing and magnitude of hydrologi-cal events to avoid negative impacts on sensitive life-history events of salmon. The latter events are sensitiveto both high and low variations in flows from the timeadult sockeye recruit to the spawning grounds in lateSeptember to the time fry complete emergence anddownstream migration in early May (Figure 2). Hydro-graphs that follow seasonal patterns approximating theall-year average generally satisfy these requirements.However, even in this case, OLRS managers appear tohave little room for error in the maintenance of flowsthat are high enough at all intervals from October toApril to meet spawning and incubation requirements. Itis also clear that hydrographs representing either extremedrought (1930–1931) or extreme flood (1996–1997) con-ditions preclude the successful management of flows toavoid either drought-and-desiccation or flood-and-scourlosses of incubating salmon eggs or alevins (Figure 2).However, there is less than a 1-in-50-year recurrence ratefor flood or drought events that are severe enough thatannual water regulation may be focused on either ofthese singular conditions. The far more common caserequires that fish and water managers shift their focusfrom managing for water conservation during an annualdry interval lasting for about 10 months (July–April) fol-lowed by a rapid change of focus to manage waterreleases to avoid excessive risk of flooding and damageto riparian property or sensitive life-history stages of fishduring the May–June freshet (Figure 2).

Information retrieved from workshops and directinterviews (Alexander et al. 2008) confirmed that annualto seasonal variations in water supply management arefurther complicated by water management objectives andconstraints associated with each of five spatial segments(Figure 3) of the OLRS. Thus, there are preferred, sea-sonal surface elevations specified for each of four valley-bottom lakes, and separate discharge ranges and/or eleva-tions to be maintained in each interconnecting segmentof the Okanagan River. Segments start at Okanagan Lakeat the head end of the valley, and end at Zosel Dam atthe outlet of Osoyoos Lake just south of the Canada–USborder (Figure 1). Seasonal preferences for lake elevationand river segment discharge ranges are dictated by sev-eral objectives including: the protection of lakeshore andriverside properties or infrastructure from flooding; mini-mization of flood or drought impacts on fisheries values;maintenance of urban and agricultural water intakes;maintenance of navigable waters and recreational activi-ties (Symonds et al. 2008; Figure 3). This wide range ofobjectives, combined with variable seasonal and annual


Dow

nloa

ded

by [

Kim

Hya

tt] a

t 09:

34 0

4 M

arch

201

5

hydrological conditions, creates a high frequency ofpotential conflicts among various groups of water usersincluding issues of whether the consideration of sockeyesalmon flow requirements at Oliver below the dam atPenticton should receive greater consideration than lakeelevation requirements to meet kokanee salmon needsupstream of the Penticton Dam at Okanagan Lake(Anonymous 1982).

To elaborate on the challenge facing resource manag-ers, Okanagan Lake receives about 80% of all surfacewater draining into the Okanagan Basin and has suffi-cient capacity to store 100% of this inflow in 1 out of 3years, and at least 66% of the inflow in 8 out of 10 years(Alexander et al. 2008). Storage during spring runoffreduces the risk of flooding and retains water for releaselater on, during lower flow periods. With a surface areaof 35,000 hectares and a preferred operating range of1.22 m, Okanagan Lake can store up to 420 million m3

of water (Anonymous 1974). This capacity is usuallysufficient to regulate preferred seasonal levels in Okana-gan, Skaha and Osoyoos Lakes, as well as the volumeand the timing of flows in interconnecting segments ofthe Okanagan River (Figure 3). However, in roughly 1in 4 years, characterized by above-average snowpack,the equivalent of 50% or more of freshet inflows must

be released to avoid flooding. By contrast, in 1 in 3years, water yields from snowpack and spring precipita-tion are small enough that storing even 100% of freshetinflows may not satisfy dry season water requirements ofboth human systems and aquatic ecosystems. The deci-sion of how much water to store at any particular time isnot an easy one. During spring freshet, the amount ofwater entering the system far exceeds the amount thatcan be released through the principal control dam at theoutlet of Okanagan Lake. Therefore, the lake must belowered several weeks before freshet to a level sufficientto store most of the freshet inflows. When inflowsexceed the volume of storage plus outflow, tens of mil-lions of dollars’ worth of real estate may be flooded (Sy-monds et al. 2008). On the other hand, if the lake isdrawn down too far prior to freshet, resultant summerwater shortages will not satisfy both irrigation and aqua-tic ecosystem needs. To further complicate matters, theBC River Forecast Centre projections of whether annual-to-seasonal variations in snowpack and spring precipita-tion will result in annual water yields that are belowaverage, average or above average are often uncertainthrough the late winter and spring interval (Figure 4)when key decisions about the timing and amount ofwater storage or release must be made.

Figure 2. Timing of sockeye salmon life-history stages (adult migration, spawning, egg/alevin incubation and fry emergence) andseasonal changes in recommended flow ranges (shaded rectangles) for the Okanagan River at Oliver under the 1982 Okanagan BasinImplementation Agreement. Seasonal hydrographs for a near-average year, the flood of record (1996–1997) and the drought of record(1930–1931) also indicate seasonal to annual intervals during which water managers must consider the impacts of drought, flood, orboth drought and flood conditions.


Dow

nloa

ded

by [

Kim

Hya

tt] a

t 09:

34 0

4 M

arch

201

5

Monitoring programs and indicators to anticipateimpending water supply issues

In the absence of highly reliable predictions of the mag-nitude and seasonal timing of water supply variationsand their impacts on potentially conflicting objectives,OLRS resource managers have depended on a variety ofnearly real-time monitoring programs and indicators toinform water storage and release decisions made at inter-vals separated by as much as several weeks (dry months)to as little as a few days (freshet months). Informationregarding temporal variations in climate (e.g. precipita-tion as rain or snow, atmospheric temperature, relativehumidity) and hydrological conditions (e.g. accumulated

snowpack, lake surface elevations, river discharge,groundwater levels) may be accessed from a network ofmonitoring stations (locations in Guy et al. 2008; Hyattand Stockwell 2013; Figure 1) established for lake, riverand stream segments in the valley bottom, and for keylocations at higher elevation (e.g. Mission Creek andBrenda Mine automated snow pillows). EnvironmentCanada and BC Ministry of Environment provide readyaccess to these types of observations at daily to hourlyintervals, from various websites. By contrast, althoughviewed as valuable, only fragmented information on sea-sonal to daily changes in the status of various life-historystages of sockeye (Hyatt and Rankin 1999) and kokanee

Figure 3. Boundaries for the five river and lake segments within the Okanagan Basin explicitly considered by the Fish and WaterManagement Tool. Bullet points summarize the focus and timing of fish and water management objectives that must be consideredwithin each geographic segment. Urban infrastructure is concentrated in cities and towns already identified (see Figure 1), while theeconomic value of this infrastructure is approximately proportional to the size of their resident populations (see Figure 1 legend).


Dow

nloa

ded

by [

Kim

Hya

tt] a

t 09:

34 0

4 M

arch

201

5

salmon (Ashley et al. 1998), occupying specific habitatswithin the OLRS, have been available from federal andprovincial fisheries personnel, at irregular intervals. Con-sequently, given decision intervals of as little as 2–3days during rapidly changing flood or drought condi-tions, uncertainties driven by natural variability, frag-mented information flow and incomplete analysis havehistorically limited the effectiveness of OLRS resourcemanagers striving to achieve an appropriate balanceamong competing objectives. However, during workshopdeliberations, a general consensus emerged among localresource managers that all concerned would benefit fromthe creation of an EDSS to facilitate the timely exchangeand synthesis of the rich information assets potentiallyavailable to improve fish and water management in theOkanagan.

Structure and process for FWMT assembly

FWMT is an internet-accessible (www.fwmt.net), multi-user EDSS for the Okanagan basin. Full documentationof the process to create this DSS as well as its contents

is available elsewhere (Alexander and Hyatt 2013), soshort overviews of both should suffice here. The overallgoal of the project was to integrate knowledge and datato provide a DSS for resource managers to make in-sea-son decisions that better balance socioeconomic, fisheriesand aquatic ecosystem values. During two design work-shops, local resource managers specified the principalOLRS management objectives to which the DSS shouldbe responsive (Table 2). The design of the resultant userinterface and outputs for FWMT were guided by theseobjectives to provide water and fisheries managers withan intuitive decision-making framework for choosingweekly water releases at Penticton Dam at the outlet ofOkanagan Lake. Release decisions, informed by inflowforecasts from the BC River Forecast Centre, are passedto FWMT’s five coupled sub-models (Figure 5) to simu-late critical performance outcomes associated with satis-fying fish and water management objectives at a varietyof lake and downriver sites (Figure 3; Table 2). The tim-ing and duration of salmon life-history stage vulnerabil-ity to water management decisions (Figure 2) aregenerally controlled by interactions among river dis-

Figure 4. Comparisons of British Columbia River Forecast Centre predicted versus observed water inflows to Okanagan Lake oneach of 1 February, 1 March, 1 April and 1 May over a 29-year interval. Forecasts provided in February and March are not givenmore weight than reference to the all-year average in making late winter to early spring management decisions, but preference isgiven to April and May forecasts, which explain enough of the observed annual discharge variance to be useful.


Dow

nloa

ded

by [

Kim

Hya

tt] a

t 09:

34 0

4 M

arch

201

5

http://www.fwmt.net

charge, lake surface elevations and seasonal water tem-perature. Thus, in addition to housing parameters andlookup information for FWMT’s biophysical sub-models,the database for the system is automatically updated eachday with actual recorded data for Okanagan Lake eleva-tion, water temperatures and Okanagan River dischargeat several sites. This “real-time” information feeds intothe hydrology and water temperature sub-models of theDSS to “self-correct” inflow forecasts and adjust sockeyeand kokanee sub-model forecasts (see details below) as acentral feature to improve confidence in FWMT modelpredictions. An internet-accessible, record-of-design doc-ument (Alexander and Hyatt 2013), managed by versionto reflect ongoing refinements, provides details of equa-tion systems, parameters, observational data and therationale underlying each sub-model. The rationale foreach sub-model is as follows.

Water supply and hydrology sub-model

The role of the water supply and hydrology sub-model(Guy et al. 2008) is to identify the relative impact ofwater management decisions on: (1) Okanagan Lake ele-vations and (2) Okanagan River discharge at severalpoints of interest. The water supply and hydrology sub-

model uses forecasts of inflows, updated with real-timehydrologic data, to determine lake elevation and riverdischarge consequences associated with alternative Pen-ticton Dam water releases. Inflow forecasts are providedby the BC River Forecast Centre, but may be replaced atan FWMT user’s discretion when supplemental observa-tions of local hydrological conditions warrant (Alexanderet al. 2008). The spatial extent of the water supply andhydrology sub-model includes the entire watershed (Fig-ure 1) of Okanagan Lake, the Okanagan River upstreamof Osoyoos Lake and the north basin of Osoyoos Lakewhere sockeye fry rear for a year following emergence.However, sub-model calculations are performed only atspecific locations within the watershed including:

� Okanagan Lake;� Okanagan River at Penticton;� Okanagan River at Okanagan Falls;� Okanagan River at McIntyre Dam; and� Okanagan River at Oliver.

Climate and water temperature sub-model

The role of the temperature sub-model (Hyatt andStockwell 2003; Hyatt et al. 2008a) is to forecast water

Table 2. Principal Okanagan Lake and River System (OLRS) management objectives identified by category during interdisciplinaryworkshops to design the Okanagan Fish and Water Management Tool (FWMT) decision support system (DSS).

Category Proposed management objective

Overall goal � Apply an ecosystem-based approach to in-season water management in the Okanagan basin� Remain consistent with the Okanagan Basin Water Agreement, but intelligently exercise flexibility that

exists

Sockeye � Increase sockeye smolt production by at least 15% or 300,000 smolts per year (relative to past watermanagement practices) through improving egg to smolt survival (incubation success and fry rearing inOsoyoos Lake)

� Manage flows to improve rearing water quality for fry in Osoyoos Lake (temperature and oxygenlevels)

� Maintain spawning gravel quality through occasional flushing flows

Kokanee � Improve shore-spawning kokanee spawning success (egg to fry survival)� Maintain Okanagan River kokanee egg to fry at acceptable levels

Flooding � Maintain flood control objectives around Okanagan Lake and along the Okanagan River, whileallowing for a reasonable balance of multi-objective risks

Water supply � Keep water intakes submerged (lower threshold for river flow and lake level)� Ensure water intakes are not harmed by high water levels (e.g. pump houses)� Ensure adequate water to meet irrigation and community/domestic demand within established standards

of lake elevation/river discharge

Navigation andrecreation

� Maintain acceptable summer water levels at boat docks and in Okanagan River for recreation,navigation and tourism


Dow

nloa

ded

by [

Kim

Hya

tt] a

t 09:

34 0

4 M

arch

201

5

temperatures at points of interest, and replace and correctthese forecasts as real-time water temperature informa-tion is made available during a water year. Water temper-atures are used principally in the kokanee and sockeyesalmon sub-models to predict the timing and duration ofkey life-history events (e.g. adult migration intervals,peak spawn dates, egg hatch and fry emergence datesand, in the case of the sockeye sub-model, fry rearingconditions in Osoyoos Lake). The water temperaturesub-model thus supplies a crucial biophysical linkage fordetermining “windows of vulnerability” to water man-agement decisions for kokanee and sockeye, as well asthe need for Osoyoos Lake flow mitigation to improvein-lake rearing conditions. The spatial extent of the tem-perature sub-model is the same as that of the water sup-ply and hydrology sub-model. However, sub-modelcalculations are performed only at:

� Okanagan Lake (Water Survey of Canada [WSC]station 08NM083);

� Okanagan River at Okanagan Falls (WSC station08NM002);

� Okanagan River at Oliver (WSC station08NM085); and

� Osoyoos Lake.

Where real-time data are not available, water temper-atures are derived from air-to-water temperature recon-structions (e.g. Hyatt and Stockwell 2003). In somecases, this involves the use of other climate (air tempera-ture) stations not listed above (see details in Hyatt et al.2008a).

Kokanee egg/alevin survival sub-model

The role of the kokanee sub-model (Andrusak et al.2008) is to identify the potential for egg/alevin mortalityof shore-spawning kokanee due to fluctuations in Okana-gan Lake levels during the egg/alevin incubation period.Kokanee populations in the Okanagan River, below Oka-nagan Lake, are not explicitly incorporated in sub-mod-els, but are assumed to benefit from downstream watermanagement decisions in a manner consistent with out-comes for sockeye salmon in the river at Oliver. Thepost-emergent fate of kokanee fry rearing in OkanaganLake is not considered to be influenced by water man-agement decisions and so is not considered in this sub-model. The spatial extent of the kokanee sub-modelincludes the entire nearshore environment of OkanaganLake. However, sub-model emergence timing calcula-tions are limited to the northwest quadrant of OkanaganLake (Figure 1). The relationships and assumptions usedfor spawning and egg-burial depth distribution as well asthe level of acceptable drawdown are best interpreted ona lake-wide average basis. The data used to developthese relationships and rules originate from multipleyears of field surveys at several index locations in Oka-nagan Lake (details in Andrusak et al. 2008).

Sockeye life-history sub-model

The role of the sockeye sub-model (Hyatt et al. 2008b)within FWMT is to identify the relative impact of waterrelease decisions on: (1) life-history event outcomes forsockeye salmon spawning in the Okanagan River and

Figure 5. The Fish and Water Management Tool (FWMT) model is a coupled set of four biophysical sub-models (1,3,4,5) of keyrelationships (among climate/hydrology, fish, water and property) that interact with a fifth, water management “rules” sub-model (2)used to predict the consequences of water management decisions for fish and other water users. FWMT software allows system usersto explore water management decision impacts by employing current data (current mode), historic data (retrospective mode) or syn-thetic, future data (prospective mode) pertaining to water supplies, climate and fish life-history states (6,7,8). Sockeye salmon smolt-to-adult return rates (SAR) are used outside of the model to anticipate future abundance changes.


Dow

nloa

ded

by [

Kim

Hya

tt] a

t 09:

34 0

4 M

arch

201

5

rearing in Osoyoos Lake and (2) sockeye fry and smoltproduction benefits that might be obtained by modifyingthese decisions. The sockeye sub-model provides aframework for capturing information about variations inwater supply years, alternative water management scenar-ios, and the consequences for sockeye productionexpressed in terms of annual fry recruitment as well assmolt production per female from a given brood year ofadults. However, the sub-model does not consider subse-quent mortality factors operating outside of the Okana-gan Basin after a given cohort of smolts migrate seawardin their second year of life.

Performance measures for a given sockeye brood-year (i.e. the calendar year of spawning) are as follows:

� Effective female spawners;� Potential egg deposition;� Sockeye egg-to-fry survival (i.e. spring-fry recruit-

ment per female);� Spring-to-fall-fry survival in Osoyoos Lake (i.e.

fall fry production per female); and� Overwinter fry survival (expressed as potential

smolts per female).

Thus, the sockeye sub-model is able to address ques-tions such as:

� What are the overall benefits of adjusting currentwater management decision rules in terms of maxi-mizing adult spawning success, fry survival orsmolt production?

� What are some of the potential bottlenecks in over-all survival that constrain the benefits of alteringwater management decision rules applied to a sin-gle life-history stage?

� How would reducing these bottlenecks throughvarious mitigation measures affect the overall pro-duction of Okanagan sockeye?

The spatial extent of the sockeye sub-model includesthe north basin of Osoyoos Lake plus upstream portionsof the Okanagan River to the point immediately belowMcIntyre Dam (Figure 1). Functional relationshipsamong river discharge, water surface elevations and hab-itat utilization by life-history stages in the sockeye sub-model are based on weighted habitat utilization valuesderived from several years of surveys focused on adultsockeye spawning in the Okanagan River, fry emergencefrom spawning beds and juvenile sockeye rearing in Os-oyoos Lake (Hourston et al. 1954; Hyatt and Rankin1999; Hyatt et al. 2003; Stockwell and Hyatt 2003).Similarly, predictions of life-history stage losses (eggs,alevins, fry) associated with river discharge and/or sur-face elevation variations are based on direct observations

of physical alteration (Summit Environmental Consul-tants Ltd. 2002) and probability estimates of continuedsurvival in the subject habitats (details in Hyatt et al.2008b).

Socioeconomic, water management “rules” sub-model

The role of the water management “rules” sub-model (Sy-monds et al. 2008) is to identify the relative impact ofOkanagan Lake water elevations and Okanagan Riverdischarge on: (1) flood-related property damage; (2) watersupply availability for agricultural irrigation and commu-nity/domestic uses; (3) navigation problems at docks andmarinas on Okanagan Lake; and (4) recreation opportuni-ties (e.g. access to beaches, river floats) in the OLRS.Accordingly, the sub-model uses outputs from the watersupply and hydrology sub-model along with supplemen-tal information (e.g. derived from engineering specifica-tions, maps, economic assessments and interviews withproperty managers, realtors and other business persons)to specify acceptable and unacceptable “rules” for lakelevels and flows at select points of interest. Spatialboundaries of the water management “rules” sub-modellargely mirror the water supply and hydrology sub-model.However, many performance measures (e.g. flood dam-age costs, irrigation water that cannot be withdrawn) areinterpreted implicitly or on an area-wide basis, instead ofbeing associated with specific locations (details in Alex-ander et al. 2008 and Symonds et al. 2008).

FWMT temporal horizon and resolution

A fundamental concept in FWMT is that of a “decisiondate.” By design, FWMT uses the best information avail-able for any particular decision date specified. A decisiondate is the specific calendar date for which a model userwishes to see a forecast of impacts from a water releasedecision. Because water management decisions andevents prior to this date are not subject to change,FWMT displays the real-time lake elevations, river flowsand water temperatures that actually occurred. The tem-poral horizon for three of FWMT’s sub-models is 1October of year n to 30 September of year n + 1 (12months). The temperature and sockeye salmon sub-mod-els (only) extend to 30 November of year n + 1 (14months) to cover life-history intervals from egg deposi-tion to smolt emigration. As an in-season managementtool, FWMT is not currently structured to automaticallyexecute multi-year simulations.

Sub-model integration

Participants in design workshops and technical meetingsemphasized the need for the clear specification of


Dow

nloa

ded

by [

Kim

Hya

tt] a

t 09:

34 0

4 M

arch

201

5

information transfers between biophysical sub-models.FWMT’s five sub-models: (1) use inputs from other sub-models and other data; (2) process management actions(only the water supply and hydrology sub-model in thiscase); (3) generate information needed by other sub-mod-els; and (4) generate performance measures used formaking decisions. The sub-model interactions required toperform these functions were identified during design in2001, by performing a “looking outward” exercise(Table 3; Hyatt et al. 2001). Sub-model experts wereasked: “What do you need to know about all the othersubsystems to predict how yours will behave?” This steprequired disciplinary specialists to consider both inputsand outputs required for their sub-models, operatingwithin the context of an entire set of linked sub-models.The exercise was intentionally interdisciplinary, andallowed the FWMT design team to clarify key linkagesthat needed to be understood in order to integrateFWMT sub-models into a single, whole-system model(Figure 5).

Management actions

The FWMT water supply and hydrology sub-modelimplements one fundamental management action: alter-native schedules of weekly water releases at PentictonDam. Other water control structures at the outlets ofSkaha and Vaseux Lakes (Figure 1) were considered butignored, as Penticton Dam usually controls OLRS dis-charge overall. However, unregulated tributary inflowsbelow Penticton Dam are incorporated, because theyconstrain management options for water releases duringthe spring freshet. As a given water year progresses,FWMT incorporates real-time hydrometric and watertemperature data to self-correct forecasts, which greatlyaids operators in the identification of near-term trendsand changing requirements to modify their proposedschedule of releases. Water release policies vary fromyear to year depending largely on the total volume anddistribution of Okanagan Lake inflow.

FWMT use and automated reports

Framework for making informed tradeoffs, balancingobjectives

FWMT users represent natural resource managers fromprivate industry, First Nations, federal and provincialinterests. The principal purpose of the DSS is to deter-mine the extent to which increases in average sockeyeproduction might be realized through improvements tothe maintenance of fish-friendly water release decisionsat Penticton Dam. This is to be accomplished throughroutine, in-season application of the DSS which links thehydrology, temperature, water management “rules” andfish production variation sub-models. While FWMT was

intended to assist with maintaining fisheries values, waterstorage and release decisions at the Penticton Dam influ-ence multiple outcomes upstream in Okanagan Lake anddownstream in the Okanagan River. Thus, the identifica-tion of the ideal water release plan to maximize sockeyeproduction in downstream areas may have collateralimpacts on potential production of kokanee fry originat-ing from shallow-water spawning beds upstream in Oka-nagan Lake. Fish-friendly decisions will also have limitsin relation to managing flood risks and water extractionsfor off-channel use throughout the OLRS. However, exe-cution of FWMT simulations, prior to key decisionpoints, provides standardized performance measures forboth fish and socioeconomic objectives to identify thepotential scope for achieving balanced water manage-ment decisions.

The workflow typically followed by FWMT softwareusers (Figure 6) is designed to allow individual resourcemanagers to quickly arrive at a set of water release rec-ommendations to satisfy their particular objectives whilerespecting the need to balance these against priorityneeds of others. The application of FWMT is facilitatedthrough access to a FWMT software user’s guide (Alex-ander et al. 2008) that provides a condensed overview ofwater management issues in the Okanagan Basin anddetailed instructions for implementing simulations. Sum-marizing these briefly here, at the point of initial login toFWMT software, a user is prompted to provide a namefor their water management “scenario” to the main entryscreen (Figure 7) and set the current decision date to beused as the starting point for their simulation. Year-spe-cific parameters (e.g. number of spawning salmon, peakspawn date, etc.) are either entered into the main entryscreen or, if unavailable, default to all-year average val-ues. The main entry screen also displays two sets ofaverage flow values for “water year” weeks extendingfrom early October of year n until the end of Septemberin year n + 1. For all weeks that have already passed,the flows displayed are the actual water releases to date,which users may not alter. For all future weeks, entriesmay be edited at a user’s discretion, after which a simu-lation run may be executed. A single simulation runrequires 2–4 minutes of elapsed time at which point anFWMT “Run Results” screen will appear. This screenprovides users with a series of graphical displays of haz-ard assessments associated with each key objective forthe five geographic segments considered within the DSS.FWMT designers recognized from the outset that thecomplexity of sub-model interactions, numeric outputand the number of potential indicators (Table 4) couldlimit the utility of FWMT for its principal users (i.e.front-line fish and water managers). To overcome theproblem of information overload and the associated timecost, the software provides a user-friendly interface thatconverts complex numeric outputs from model simula-


Dow

nloa

ded

by [

Kim

Hya

tt] a

t 09:

34 0

4 M

arch

201

5

tions into key performance indicators (e.g. sockeye eggor fry losses, dollar value of insurance claims for flooddamage, etc.).

FWMT performance indicators are expressed in agraphical form that follows a familiar “traffic light” prin-ciple (green = go ahead; amber = exercise caution; red =stop or risk certain damage). The graphical user interface(GUI) and traffic light indicators largely eliminaterequirements for managers to identify precise numericoutcomes to achieve prudent water managementdecisions. For example, the hazard report display for

Okanagan Lake (Figure 8a) upstream of the main damidentifies the state of risk indicators for flood control,domestic intakes, navigation and survival of kokanee sal-mon eggs–alevins resulting from the set of weekly waterreleases specified by the user. By contrast, the hazardreport display at Oliver (Figure 8b), roughly 70 kmdownstream of Penticton Dam, provides information notonly on flood control and water intake indicators but alsofor indicators of flow-induced risk to sockeye salmoneggs and alevins. Moreover, the rapidity with whichsimulations may be completed allows FWMT users to

Table 3. Fish and Water Management Tool (FWMT) “looking outward” matrix. Each element in the matrix represents a transfer ofinformation between subsystems, from a row to a column.

From ⇩ To ➩Dam operations,hydrology, socioeconomic Kokanee Sockeye

Dam operations,hydrology,socioeconomic

� Levers/limits/triggersdriving rules forOkanagan Lakeoperations to deal withinflow variability

� Forecast Okanagan Lakelevels on 1October (want thisas low as possible)

� Actual daily average lakelevels from 1 Octoberthrough to end of incubationperiod

� Forecast weekly OkanaganRiver flow at Oliver

� Maximum expected averagedaily flow at Oliver in eachweek through to end ofincubation period

Temperature � Forecast (and, as available,actual) daily lake watertemperatures from date ofpeak spawning through totypical end of emergenceperiod

� Forecast (and as available,actual) daily water temperaturesin the Okanagan River nearOliver in weeks just prior tospawning through to the end oftypical emergence period

� Forecast (and, as available,actual) daily water temperaturesin the north basin of OsoyoosLake from sockeye emergencethrough to 30 November

Kokanee � Egg burial depth relations� Rules for time of peak spawn� Rules for emergence timing

based on real-time andforecast water temperatures

� Impact of lake decline onkokanee eggs

Sockeye � Rules for time of peak spawn� Rules for emergence timing

based on real-time and forecastwater temperatures

� Egg mortality associated withgiven Okanagan River flows(scour and desiccation events)

� Density-dependent andindependent survivalrelationships for fry rearing inOsoyoos Lake

Actions � Alternative weeklyOkanagan Lake damrelease schedules


Dow

nloa

ded

by [

Kim

Hya

tt] a

t 09:

34 0

4 M

arch

201

5

efficiently execute several scenario simulations toimprove the precision of “solution space” identificationprior to a given water release decision. A given usermay also share water management scenarios, online, withother users who can then copy, modify and refine oneanother’s suggested water release schedule for OkanaganLake Dam.

For example, during the 2009–2010 fish and wateryear, a warm and dry late summer interval emerged aspotentially problematic with respect to fully meetingwater extraction, recreation and fish habitat maintenanceobjectives in the north basin of Osoyoos Lake. Succes-sive FWMT simulations supported the rapid assembly ofa hazard indicators “report card” based on alternate waterrelease scenarios (Figure 9). This report suggested a needto increase water releases, from 10.7 m3/s to either 12.7or even 18.3 m3/s, to achieve a better balance of ecosys-tem and human system maintenance. However, watermanagers remained reluctant to increase discharge at thistime because it would result in Okanagan Lake levelsdropping well below the preferred October target eleva-tion. Low October elevations pose a threat to kokaneespawning and increase the risk of winter dewatering ofeggs–alevins of kokanee in Okanagan Lake and sockeye

downstream in the Okanagan River. Consequently, theFWMT operations team recommended enhanced moni-toring of the observed (as opposed to predicted) status ofjuvenile and adult sockeye salmon in Osoyoos Lake withthe understanding that at the first signs of elevated mor-talities this decision would be reconsidered. In the end,late summer environmental conditions moderated suffi-ciently to avoid increasing weekly water releases. Never-theless, the application of FWMT increased theconsideration of fish-friendly maintenance requirementscommonly neglected during the pre-FWMT deploymentinterval.

Impact of FWMT on water management compliancewith OBIA guidelines

Because OLRS water storage and release decisions areultimately the responsibility of BC water managementpersonnel, the extent to which the deployment of FWMTwould alter levels of compliance with OBIA fisheriesguidelines for maintenance of river flows and lake eleva-tions was assessed. Consequently, in each year betweenthe fall of 2003 and the spring of 2014, a detailed recordof water management decisions and outcomes (“test”

Figure 6. A summary of the work flow for Fish and Water Management Tool software users (Alexander et al. 2008).


Dow

nloa

ded

by [

Kim

Hya

tt] a

t 09:

34 0

4 M

arch

201

5

observations) associated with FWMT use by the multi-party operations team was assessed (Hyatt and Bull2007; Hyatt et al. 2009; Hyatt and Stockwell 2010; Hy-att and Stockwell 2013). An extended record of observa-tions of the frequency and degree of non-compliant (NC)drawdowns of Okanagan Lake (affecting kokanee) andOkanagan River flows (affecting sockeye) during fourmanagement intervals was assembled (Figure 10) includ-

ing: a pre-OBIA guidelines interval (1967–1976), anFWMT “control-years” interval (i.e. C-FWMT, 1977–1998 when OBIA guidelines were in force, but FWMTwas not in use); a build-FWMT interval (B-FWMT,1999–2002) and a use-FWMT interval (U-FWMT, 2003–2013).

Lake elevation changes, between the time of fall eggdeposition and spring fry emergence by kokanee salmon,

Figure 7. Modified version of a screensaver shot of the main entry form that allows users to name a scenario, enter year-specificparameters for kokanee and sockeye salmon, propose water releases at weekly intervals for the remainder of a given water year (wateryear starts 1 September of year n and ends 31 August of year n + 1). Users have discretion over sharing scenario inputs and out-comes with other system users.


Dow

nloa

ded

by [

Kim

Hya

tt] a

t 09:

34 0

4 M

arch

201

5

Table

4.Sum

maryof

keyevents

oractiv

ities,risk

factorsor

processes,andindicators

ofrelevanceforusersof

theOkanaganFishandWater

Managem

entToo

l(FWMT)deci-

sion

supp

ortsystem

.The

majority

ofindicators

aregeneratedwith

intheFWMTsystem

asmod

elpredictio

ns(P)or

measuredob

servations

(O)that

areim

ported

innear

real-tim

e.Stars

(❖)representtheprim

aryindicators

used

foron

goingevaluatio

nof

changesin

risk.Circles

(o)representasubset

ofdetailedindicators.A

smallersetof

supp

lementalindi-

cators

aregeneratedandaccessed

byusersfrom

outsideof

theFWMTapplication(see

text

forfurtherexplanation).

Eventor

activity

Socioecon

omic

Pressure,statusan

dtrendindicator(s)

Sou

rce

Riskfactor(s)or

process

Inside

FW

MT

Outside

FW

MT

PO

PO

Okan

agan

Lak

eat

Kelow

na

OkanaganLake

floo

ding

and

associated

prop

erty

damage

❖Okanagandaily

toseason

alsnow

pack

values

relativ

eto

average

XX

Surface

elevation

ofOkanagan

Lake

OkanaganLake

drou

ghtandstorage

deficitthat

impacts

water

access

for

irrigatio

nanddo

mestic

water

intakes

❖BC

River

ForecastCentreor

user-specified

water

supp

lyforecast

XX

oRegionalsnow

pack

and/or

rainfalleventsrelativ

eto

average

Xo

Okanagandaily

tomon

thly

rainfallvalues

relativ

eto

average

XX

oHou

rlyto

daily

inflow

sfrom

Mission

Creek

toOkanaganLake

XX

oPentictonDam

flow

releases

byho

urX

Xo

OkanaganRiver

dischargeby

hour

XX

❖Net

inflow

sfrom

tributariesrelativ

eto

weeklyor

mon

thly

average

XX

❖OkanaganLakedaily

toweeklylake

levelrelativ

eto

season

altargets

XX

Okan

agan

River

atPenticton

Flood

-ind

uced

damage

toPentictonchannel

and/or

floo

ding

and

water

infiltrationof

riparian

prop

erties

oPentictonDam

flow

releases

byho

urX

X

Discharge

and

water

levelin

Penticton

channel

Droug

ht-ind

uced

expo

sure

ofdo

mestic

andirrigatio

nwater

intakes

oOkanaganRiver

dischargeby

hour

XX

oInflow

sfrom

tributariesrelativ

eto

weeklyor

mon

thly

average

XX

oRiparianland

ownercommentaries

re:specificim

pactson

prop

erty

Xo

Flowsin

Pentictonchannelwith

inrang

eforrecreatio

nal“tub

ing”

XX

❖OkanaganRiver

daily

dischargerelativ

eto

season

altargets

XX

Okan

agan

River

atOliv

erFlood

-ind

uced

damage

tochannelat

Oliv

erand/or

floo

ding

and

water

infiltrationof

riparian

prop

erties

oPentictonDam

flow

releases

byho

urX

X

Discharge

and

water

levelin

Oliv

erchannel

Droug

ht-ind

uced

expo

sure

ofdo

mestic

andmajor

irrigatio

nintake

atMcIntyre

Dam

oOkanaganRiver

dischargeby

hour

XX

oInflow

sfrom

tributariesrelativ

eto

weeklyor

mon

thly

average

XX

oRiparianland

ownercommentaries

re:specificim

pactson

prop

erty

X❖

OkanaganRiver

daily

dischargerelativ

eto

season

altargets

XX

(Con

tinued)


Dow

nloa

ded

by [

Kim

Hya

tt] a

t 09:

34 0

4 M

arch

201

5

Table4.

(Con

tinued).

Eventor

activity

Socioecon

omic

Pressure,statusan

dtrendindicator(s)

Sou

rce

Riskfactor(s)or

process

Inside

FW

MT

Outside

FW

MT

Ecological

PO

PO

Kok

anee

Salmon

inOkan

agan

Lak

eat

Kelow

na

Riskof

egg/alevin

desiccationandloss

dueto

spaw

ndepth

andsubsequent

lake

leveldraw

down

betweentim

eof

egg

depo

sitio

nandfry

emergence

oNo.

ofspaw

ners

bylake

area

(SE,NE,NW)

X

Kok

anee

spaw

ning

andincubatio

nsuccesson

OkanaganLake

beaches(SE,

NE,NW)

oSpawndepth

Xo

Lakelevel

XX

oIncubatio

ntemperature

andaccumulated

thermal

units

(ATUs)

XX

oEgg

hatchandfryem

ergencedate

X❖

Magnitude

ofdraw

down-inducedegg/alevin

loss

during

incubatio

n

SockeyeSalmon

intheOkan

agan

River

atOliv

erAdu

ltsalm

onaccess

tospaw

ning

area(s)

Migratio

nblockage

atvertical

drop

-structures

dueto

high

discharge;

access

tospaw

ning

habitatredu

ceddu

eto

low

discharge

oNo.

ofadultsockeyein

riverine

spaw

ning

grou

nds

XX

oNo.

ofadultsockeyein

specificspaw

ning

areasandhabitats

XX

❖Discharge

relativ

eto

migratio

nandspaw

ning

compliancerang

eX

X

Egg

/alevin

incubatio

nand

fryem

ergence

success

Flood

ordrou

ght

impactson

egg/alevin

incubatio

nandfry

emergencesuccess

oOkanagandaily

toseason

alsnow

-packvalues

relativ

eto

average

XX

oOkanagandaily

tomon

thly

rainfallvalues

relativ

eto

average

XX

oOkanagandaily

toweeklylake

levelrelativ

eto

average

XX

oPentictonDam

flow

releases

byho

urX

Xo

OkanaganRiver

dischargeby

hour

XX

oUnregulated

tributarydischargeby

hour

XX

oOkanaganRiver

incubatio

ntemperature

andATUs

XX

oEgg

hatchdates

XX

❖Scour

anddesiccationevent-ov

er-thresho

lddrivers

XX

❖Fry

emergencedates

XX

❖Early

summer

fryrecruitm

entindex(no.

spaw

ner–1)to

Osoyo

osLake

XX

SockeyeSalmon

inOsoyo

osLak

eFry

recruitm

entto

Osoyo

osLake

Flood

ordrou

ghtwater

levelor

flow

impacts

onfrymigratio

nor

emergencesuccess

❖Discharge

ofOkanaganRiver

atOliv

errelativ

eto

emergenceandmigratio

ncompliancerang

eX

X❖

Early

summer

fryrecruitm

entindex(fry

spaw

ner–1)to

Osoyo

osL.

XX


Dow

nloa

ded

by [

Kim

Hya

tt] a

t 09:

34 0

4 M

arch

201

5

Figure 8. (a) Okanagan Lake segment of screen display presented to Fish and Water Management Tool (FWMT) users upon openingthe multiple objective hazard assessment (MOHA) report. Water management performance indicators informing decisions for Okana-gan Lake include: foreshore flood control, domestic water intake integrity, launching ramp and dock integrity, navigation ease andkokanee egg/alevin risk of desiccation. The solid black line identifies lake elevations to date. The solid blue line identifies simulated,lake-level outcomes associated with a proposed set of water releases. Green diamonds are preferred benchmarks for maintenance ofseasonal water levels as defined by the Okanagan Basin Implementation Agreement (OBIA). The yellow triangle in late April is analternate lake-level benchmark (as defined by the Okanagan Basin Agreement [OBA]) given certainty of high freshet and flood risk iflate April snowpack is far above average. The yellow triangle in late October is the preferred lake elevation benchmark for fallspawning by kokanee on Okanagan Lake beaches. Green, amber and red segments of indicator bars follow low, moderate and highrisk classifications of lake elevations associated with failure to achieve performance objectives tracked via specific indicators. The hor-izontal length of each bar reflects the interval during which a given, and sometimes competing, objective must be met. (b) The screendisplay associated with the MOHA report for a segment of the Okanagan River at Oliver. Flow regulation at Oliver is critical to meet-ing seasonal requirements for several life-history stages of sockeye adults, eggs, alevins and emergent fry (see text for details). Yellowtriangles identify the position of the threshold above which river discharge will induce losses of sockeye eggs and alevins to flood-and-scour losses. Red squares indicate the position of the threshold below which river discharge will induce losses of eggs/alevins todrought and desiccation. Line and indicator bars follow the conventions described above. Dashed lines (black and blue) indicateobserved and projected discharges, respectively, originating from unregulated tributaries downstream of the Okanagan Lake Dam.Unregulated tributary discharge may require 50–75% of flood-channel capacity during freshets and thus constrain major water releasesfrom Penticton Dam for periods of several weeks.


Dow

nloa

ded

by [

Kim

Hya

tt] a

t 09:

34 0

4 M

arch

201

5

determine annual proportions of eggs–alevins vulnerableto desiccation or stranding losses in shallow beach-spawning areas near Kelowna. The all-year averagedrawdowns of Okanagan Lake were statistically indistin-guishable between pre-OBIA and OBIA intervalsbetween 1967 and 1998 (Figure 10a), suggesting thatOBIA guidelines had little impact on this element ofwater management. By contrast, average lake-level draw-downs were significantly lower after 1998 during boththe B-FWMT and U-FWMT intervals, suggesting thatthe development and use of FWMT improved watermanagement compliance with OBIA fisheries guidelines.Specifically, OBIA guidelines to protect eggs–alevins ofkokanee salmon suggest maximum drawdowns of 15 cmof Okanagan Lake between the time of egg deposition inmid-October and egg hatch by 1 February, and then 35cm by the end of February (Hyatt and Bull 2007). Anexamination of the frequency of lake drawdowns duringthese seasonal intervals indicates significantly higher fre-quencies of NC-drawdowns of 15 cm by 1 February dur-ing pre-FWMT “control years” versus all FWMT years(data not shown), and of 35-cm drawdowns between pre-

OBIA versus all OBIA years (i.e. 1976–2013) by the endof February (Figure 10a, Chi square, corrected for conti-nuity, alpha = 0.01, df = 1 in both cases). Differencesfrom comparisons among all other intervals were not sta-tistically significant.

OBIA guidelines to protect incubating eggs–alevinsof sockeye salmon in the Okanagan River near Oliversuggest maximum flows of < 28.3 m3/s to avoid scourand > 5 m3/s to avoid desiccation impacts between thetime of egg deposition in mid-October and the time offry emergence at the end of April (Table 1). An exami-nation of Okanagan River seasonal flows revealed thatthe mean number of NC-flow days did not differ signifi-cantly between pre-OBIA and OBIA intervals between1967 and 1998 (Figure 10b), suggesting again that OBIAguidelines had little impact on routinely achieving fish-friendly flows to protect incubating sockeye eggs–alev-ins. However, the average number of NC-flow daysdeclined precipitously after 1998, coincident withFWMT development and use. Further examination indi-cated that NC-flow days, as a proportion of total days ofegg–alevin incubation, differed significantly such that

Figure 9. Summary of hazard indicator reports derived from three Fish and Water Management Tool (FWMT) scenario simulations(FWMT-569, FWMT-561 and FWMT-568) executed in early August 2009 (Hyatt and Stockwell, unpublished observations).“Squeeze” under FWMT-568 refers to extreme reductions in the volume of water in Osoyoos Lake exhibiting temperature and oxygenvalues preferred by sockeye fry for rearing or by adults for pre-spawn holding.


Dow

nloa

ded

by [

Kim

Hya

tt] a

t 09:

34 0

4 M

arch

201

5

pre-OBIA > C-FWMT > B-FWMT and U-FWMT forNC-flow days (alpha = 0.01, one-tailed test; Zar 1984,395). Proportions of NC-flow days did not differ signifi-cantly between the B-FWMT and U-FWMT intervals.

Although naturally occurring differences in hydrologybetween “management” intervals might potentiallyaccount for differences in NC-flow outcomes, this pros-pect was minimized by restricting test observations to

Figure 10. Summary observations from a before-and-after impact assessment of Fish and Water Management Tool (FWMT) deploy-ment on compliance with fish friendly guidelines by fish and water managers during consecutive sets of years defined by: the absenceof Okanagan Basin Implementation Agreement (OBIA) fisheries guidelines (pre-OBIA, 1967–1975), OBIA guidelines in force(1976–2013), where the latter then include pre-FWMT “control-years” (C-FWMT, 1976–1998), FWMT design-and-build years (B-FWMT, 1999–2002) and years of active FWMT use (U-FWMT, 2003–2013). Observations restricted to similar water supply yearswere used to test for differences between the means of lake level drawdowns, and the means or frequencies of non-compliant lakelevels and river flows (see text for details). The years 1969, 1971, 1973, 1982, 1995 and 1996 were excluded from statistical compar-isons because they exhibited subject interval hydrographs outside of the 190–750 million cubic meter seasonal range common to theremaining pre-FWMT and FWMT years. (a) Annual decline of Okanagan Lake level between 15 October and 28 February for eachmanagement interval defined above. (b) Number of days of non-compliant (NC) flows in the Okanagan River at Oliver. Non-compli-ance of observed discharge is classified according to the degree of departure from either an upper flood-and-scour threshold (beginsat > 28.3 m3/s) or a drought-and-desiccation threshold (begins at < 5 m3s) for sockeye redds. All-year means (x) and 95% confidenceintervals (CIs) of compliance indicators are shown for each management interval in (a) and (b).


Dow

nloa

ded

by [

Kim

Hya

tt] a

t 09:

34 0

4 M

arch

201

5

years of similar hydrology among subject intervals (i.e.similar frequency and nearly identical range of cumula-tive discharges within the main test and control-yearintervals; see Figure 10b).

These results strongly support an inference that,despite good intentions, water managers were frequentlyunable to comply with OBIA guidelines prior to 1999. Bycontrast, beginning in 1999, the development and use ofFWMT by an interdisciplinary team of fish and watermanagers contributed in major ways to significantimprovements in the compliance of water managementdecisions with fish-friendly guidelines affecting kokaneesalmon in Okanagan Lake and sockeye salmon in the Oka-nagan River on a daily basis. However, even short inter-vals of NC-flows have the potential to scour or desiccatesockeye eggs and alevins. Consequently, the frequency ofyears in which compliant flows were maintained over theentire incubation interval is of greater biological relevance.The application of Fisher’s exact test (Zar 1984, 391) indi-cates that the number of NC-flow years was significantlygreater during the pre-FWMT “control” interval (1977–1998) than during the all-FWMT (1999–2013) orU-FWMT (2003–2013) intervals (alpha = 0.05 in bothcases). Differences among all other intervals were not sta-tistically significant at this annual scale (Figure 10b).There were clearly too few years (n = 4) to establishimproved flow compliance during the B-FWMT phasewith certainty. However, the apparent improvement isbelieved to be real, even during this short interval, and canbe attributed to: more timely provision of incoming fieldobservations, supplemental analysis–synthesis of informa-tion on the status of fish and water indicators, and elevatedlevels of interagency dialogue supported by the projectteam during the B-FWMT phase.

Discussion

Results from more than a decade of operational testinghave demonstrated a wide range of benefits surroundingthe development and deployment of FWMT. First andforemost, this DSS has provided a solution to the previ-ous inability of water managers to effectively use the fullrange of quantitative relationships and incoming informa-tion available to satisfy competing fish and water man-agement objectives when time intervals for affectingbalanced tradeoff decisions were as short as a few days(e.g. during spring freshet). FWMT deployment has alsofacilitated an elevated level of ongoing, multi-partyengagement in the regulation of OLRS water supplies toachieve an improved balance of outcomes regardingcompeting objectives to meet both human system andnatural system requirements (Table 2). Prior to FWMTdeployment, various parties (e.g. BC Ministry of Envi-ronment, Fisheries and Oceans Canada [DFO], the Oka-nagan Nation Alliance, US Fish and Wildlife Service,

etc.) frequently made recommendations from a singleperspective, without any in-depth understanding of themulti-objective tradeoffs entailed in adopting a givenposition. However, FWMT-mediated engagement hashelped educate water management engineers about theneeds of threatened fish populations, and enabled fisher-ies biologists to gain a deeper understanding of floodmanagement, multi-faceted water supply and recreationalobjectives.

Another valuable feature of FWMT is its contributionto overcoming collaboration barriers imposed by the dis-persed locations of resource agency advisers (e.g. Pentic-ton, Kamloops, Kelowna, Vancouver, etc.) and frequentturnovers in personnel. Without FWMT, these organiza-tional factors remained a hindrance to improvements inwater management performance. Consequently, compli-ance with OBIA fisheries guidelines was low (Figure 10)and interactions among resource managers were oftenadversarial. By contrast, effective use and interpretationof outputs from FWMT exposed decision makers, regard-less of their geographic location, to a broader range ofinformation and outcomes associated with a given waterregulation decision, and encouraged them to identifywater release options to satisfy their own as well as oth-ers’ interests. Finally, concerted efforts expended duringthe design and development stages of FWMT capturedand incorporated forms of corporate knowledge havingeither partial or even no documentation (e.g. “if–thendecision rules” for OLRS maintenance, relative priorityof objectives, etc.).

Although there are many benefits to the developmentand application of FWMT, there are also limitations.Development and testing of this DSS required resourcesabove and beyond those generally available in Canadaincluding: a rich foundation of biophysical and regulatoryinformation, the extended attention of an interdisciplinaryteam (> 10 years) and significant, cumulative expendi-tures for initial data collection and development (> CAD$500,000) followed by a decade of monitoring and test-ing (> $1,000,000). The unique bilateral history of fed-eral, provincial and state government involvement(s) inthe development of fish and water management in theOLRS created an extensive knowledge base on which tobuild this EDSS. Indeed, the pioneering work of Cana-dian and US authorities to provide an empirical and theo-retical basis to specify “fish friendly” flows in terms ofweighted habitat-use functions focused on separate life-history stages of salmonids (Hourston et al. 1954) pre-dates documentation of the similar, and widely used, in-stream flow incremental methodology (IFIM; Bovee et al.1982; Stalnaker et al. 1995) by at least 30 years.

Even with this advantage, FWMT is still focused onthe use of “sentinel salmonids” as surrogate indicatorsfor the protection of a broader range of biodiversity andecosystem integrity values now challenging managers of


Dow

nloa

ded

by [

Kim

Hya

tt] a

t 09:

34 0

4 M

arch

201

5

regulated lake and river systems. This is likely to remainthe case for the foreseeable future, given the highly frag-mented knowledge of cause-and-effect relationshipsbetween aquatic communities and their associated flowregimes, which together limit their direct utility forwatershed management decisions (Bradford 2008). How-ever, there are practical steps that may be taken tobroaden the range of sentinel species requirements incor-porated within FWMT to increase its scope to functionas an effective EDSS. For example, Rocky MountainRidged Mussel (RMRM, Gonidea angulata) is listed asan endangered species under Canada’s Species at RiskAct, in part because their distribution in Canada is lim-ited to only the Okanagan Basin. Because RMRM colo-nies occupy shallow beach areas in Okanagan Lake anda few restricted segments of the Okanagan River, theyare vulnerable to changes in either lake surface elevationor river discharge falling outside of a specified range.Accordingly, in 2013, a DFO study team added perfor-mance indicators to FWMT to provide timely informa-tion to regulatory authorities of any imminent threats ofpotential losses of RMRM associated with natural hydro-logical changes and/or water regulation decisions (Hyattet al. 2013). This was readily accomplished because ofthe modular design and construction of sub-models thatdrive FWMT outputs. Further additions are feasible toaddress the needs of a broader range of aquatic biota asnew knowledge is accumulated to warrant their inclu-sion.

Conclusions

The 1982 Canada–BC Okanagan Basin ImplementationAgreement specified preferred seasonal lake levels andranges of regulated discharge in the Okanagan Lake andRiver System to protect aquatic ecosystems and fisheriesvalues without unduly compromising water managementto satisfy human system needs, but historical complianceto meet fisheries needs was low prior to 1998.

Difficulties in maintaining compliance with the “fishfriendly” flow recommendations of the OBIA wererelated to the complexity of balancing fisheries, floodcontrol and water allocation benefits through the year,given large uncertainties in: (1) forecasts of annual andseasonal water supplies; (2) the exact timing of salmonlife-history events that control their vulnerability in aparticular year to losses from flood-and-scour ordrought-and-desiccation processes; (3) the magnitude offish losses likely to be caused by deviations from recom-mended lake-level or river-flow ranges; and (4) simulta-neous risks of “significant” property losses associatedwith seasonal maintenance of fish-friendly lake elevationand river discharge levels given either flood or droughtevents.

Decision intervals of as little as 2–3 days during rap-idly changing flood or drought conditions, uncertaintiesdriven by natural variability, fragmented informationflow, data inconsistencies and incomplete analysis havehistorically limited the effectiveness of OLRS resourcemanagers striving to achieve an appropriate balanceamong competing objectives.

Beginning in 1998, a multi-party team drawn fromindustry, First Nations and government agencies began acollaborative effort to design, build and deploy an envi-ronmental decision support system to enable resourcemanagers to achieve a better balance in managing waterresources to meet both human and natural system needs.

The Fish and Water Management Tool decision sup-port system (FWMT) provides a near real-time riskassessment framework consisting of a coupled set of fourbiophysical models of key relationships (among climate,hydrology, fish, water and “property”) that interact witha fifth water management “rules” sub-model used to pre-dict the consequences of water management decisionsfor fish and other water users. The FWMT softwareallows system users to explore water management deci-sion impacts by employing current data, historic data orsynthetic future data to assess complex, cause-and-effectinteractions among water supplies, climate and risks toproperty, developed infrastructure, recreational activitiesand fish life-history outcomes.

Comparisons of pre-FWMT “control” versus FWMT-deployment years of observations (n = 20 and 11,respectively) indicate statistically significant improve-ments at both daily and annual time scales in levels ofcompliance of water management decisions with OBIAfisheries guidelines for sockeye and kokanee salmon dur-ing critical egg-to-fry emergence stages.

The development and deployment of FWMT havebeen a success to date in that they: (1) have providedadditional, useful information for decision making, (2)reduced the time and money required for fish and watermanagement-plan formulation and evaluation, and (3)have facilitated the implementation of management solu-tions that otherwise would not have been identified.

Development and especially rigorous testing of EDS-Ss to facilitate management for biodiversity values orecosystem integrity are relatively rare (see Zhang et al.2014), but they offer an as-yet underutilized approach byaquatic ecologists and water managers to tackle problemsof complex and uncertain systems where high-valueresource conflicts may be chronic.

AcknowledgementsThe decade-long process of developing and applying FWMThas produced significant technical and cognitive advances infish and water management in the basin. These advances wouldnot have been possible without the unparalleled collaboration


Dow

nloa

ded

by [

Kim

Hya

tt] a

t 09:

34 0

4 M

arch

201

5

sustained by the three “parties” (Fisheries and Oceans Canada;BC Forests, Lands, and Natural Resource Operations; Okana-gan Nation Alliance) of the Canadian Okanagan Basin Techni-cal Working Group and their partners. Rick Klinge and TomKahler, on behalf of Public Utility District No. 1 of DouglasCounty, Washington, provided funding and encouragement thatwas vital to all phases of FWMT development and testing. Wealso thank the several BC water managers who have put bothFWMT and the rest of us through our paces while teaching usthe “art” of complex fish and water management. Brian Sy-monds, in particular, was especially generous with his knowl-edge of OLRS history and management practices, and hisunwavering support.

ReferencesAhmad, S. and S. P. Simonovic. 2006. An intelligent decision

support system for management of floods. Water ResourcesManagement 20(3): 391–410.

Alexander, C. A. D., and Hyatt, K. D., ed. 2013. The Okana-gan Fish-and-Water Management Tool (Ok-FWMT) recordof design (v. 2.4.000). Report prepared for the CanadianOkanagan Basin Technical Working Group (COBTWG)and Douglas County Public Utility District No. 1(DCPUD). Draft report submitted to COBTWG andDCPUD.

Alexander, C. A. D., Hyatt, K., and B. Symonds. 2008. TheOkanagan Fish/Water Management Tool: Guidelines forapprentice water managers (v.2.1.000). Report prepared forthe Canadian Okanagan Basin Technical Working Group,Kamloops, BC and Douglas County Public Utility District,Wenatchee, WA.

Andrusak, H., A. Wilson, and S. Matthews. 2008. The kokaneesubmodel of the Okanagan Fish/Water Management Tool(FWMT). In The Okanagan/Fish Water Management(OFWM) Tool: Record of design (2.4.000), ed. C. A. D.Alexander and K. D. Hyatt, 77–96. 2013. Report preparedfor Fisheries and Oceans Canada, Vancouver, BC andDouglas County Public Utility District, East Wenatchee,WA.

Anonymous. 1973. Pacific salmon population and habitatrequirements. Task 162. Report for the Okanagan StudyCommittee, Canada–British Columbia Okanagan BasinAgreement.

Anonymous. 1974. Main report of the Consultative Boardincluding the Comprehensive Framework Plan. Preparedunder the Canada–British Columbia Okanagan BasinAgreement.

Anonymous. 1982. Report on the Okanagan Basin Implementa-tion Agreement (OBIA). Okanagan Basin ImplementationBoard Canada.

Ashley, K., B. Shepherd, D. Sebastian, L. Thompson, L. Vid-manic, P. Ward, H.A. Yassien, et al. 1998. Okanagan LakeAction Plan year 1 (1996–97) and year 2 (1997–98)Report. Fisheries Project Report No. RD 73. Province ofBritish Columbia, Ministry of Fisheries, Fisheries Manage-ment Branch. Penticton: BC.

Bovee, K. D. 1982. A guide to stream habitat analysis usingthe instream incremental methodology. United StatesDepartment of Interior Fish and Wildlife Service InstreamInformation Flow Paper 12. FWS/OBS-82/26.

Bradford, A. 2008. An ecological flow assessment framework:Building a bridge to implementation in Canada. CanadianWater Resources Journal 33(3): 215–232.

Bull, C. J. 1999. Fisheries habitat in the Okanagan Riverphase 2: Investigation of selected options. Report preparedfor Public Utility District of Douglas County, WA.

Dobson, D. 2004. Hydrology and watershed management.Chap. 13 in Okanagan geology, British Columbia. 2nd ed.,ed. M.A. Roed and J.D. Greenough, Kelowna GeologyCommittee. Kelowna, BC: Sandhill Book Marketing.

Guy, B., L. Uunila, B. Philips, and C. Alexander. 2008. Thewater supply/hydrology submodel of the Okanagan Fishand Water Management Tool. In The Okanagan Fish-and-Water Management Tool (Ok-FWMT): Record of design(v. 2.4.000), ed. C. Alexander and K. Hyatt, 19–48. 2013.Report prepared for Fisheries and Oceans Canada, Vancou-ver, BC and Douglas County Public Utility District, EastWenatchee, WA.

Hourston, W. B., C. H. Clay, E. W. Burridge, F. C. Lucas, D.R. Johnson, H. T. Heg, W. R. McKinley et al. 1954. Thesalmon problems associated with the proposed flood con-trol project on the Okanagan River, British Columbia, Can-ada. United States Fish and Wildlife Service, WashingtonState Department of Fisheries and Canada Department ofFisheries.

Hyatt, K. D. and C. Bull. 2007. Fish and Water ManagementTool project assessments: Record of management strategyand decisions for 2005. Canadian Manuscript Report ofFisheries and Aquatic Sciences 2808.

Hyatt K. D. and D. P. Rankin. 1999. A habitat based evalua-tion of Okanagan sockeye salmon objectives. PSARCWorking Paper S99-18. Canada Department of Fisheriesand Oceans, Pacific Science Advisory Review Committee,Pacific Biological Station, Nanaimo: BC. http://www.dfo-mpo.gc.ca/csas/Csas/publications/ResDocs-DocRech/1999/1999_191_e.htm (accessed September, 2014).

Hyatt, K. D. and M. M. Stockwell. 2003. Analysis of seasonalthermal regimes of selected aquatic habitats for salmonidpopulations of interest to the Okanagan Fish and WaterManagement Tools (FWMT) project. Canadian ManuscriptReport of Fisheries and Aquatic Sciences 2618.

Hyatt, K. D. and M. M. Stockwell. 2010. Fish and Water Man-agement Tool Project Assessments: Record of managementstrategy and decisions for the 2006–2007 water year.Canadian Manuscript Report of Fisheries and Aquatic Sci-ences 2913.

Hyatt, K. D. and M. M. Stockwell. 2013. Fish and Water Man-agement Tool project assessments: Record of managementstrategy and decisions for the 2007–2008 water year.Canadian Manuscript Report of Fisheries and Aquatic Sci-ences 3022.

Hyatt, K. D., C. Bull, and M. M. Stockwell. 2009. OkanaganFish and Water Management Tool Project assessments:Record of management strategy and decisions for 2005–2006 Fish-and-Water Year. Canadian Manuscript Report ofFisheries and Aquatic Sciences 2897.

Hyatt, K. D., F. Poulsen, and C. A. D. Alexander. 2013. TheRocky Mountain Ridged Mussel submodel of the Okana-gan Fish/Water Management (OKFWM) Tool. In TheOkanagan/Fish Water Management (OKFWM): Recordof design (v. 2.4.000), ed. C. A. D. Alexander andK. D. Hyatt, 54–76. 2013. Report prepared for Fisheriesand Oceans Canada, Vancouver, BC and Douglas CountyPublic Utility District, East Wenatchee, WA.

Hyatt, K. D., M. M. Stockwell, and D. P. Rankin. 2003. Impactand adaptation responses of Okanagan River sockeye sal-mon (Oncorhynchus nerka) to climate variation and changeeffects during freshwater migration: stock restoration and


Dow

nloa

ded

by [

Kim

Hya

tt] a

t 09:

34 0

4 M

arch

201

5

http://www.dfo-mpo.gc.ca/csas/Csas/publications/ResDocs-DocRech/1999/1999_191_e.htm



fisheries management implications. Canadian WaterResources Journal 28(4): 689–713.

Hyatt, K., M. Stockwell, H. Stiff, and C. A. D. Alexander.2008a. The temperature sub-model of the Okanagan Fish-and-Water Management Tool (Ok-FWMT). In The Okana-gan/Fish Water Management (OKFWM) Tool: Record ofdesign (v. 2.4.000), ed. C. A. D. Alexander and K. D. Hy-att, 49–71. 2013. Report prepared for Fisheries and OceansCanada, Vancouver, BC and Douglas County Public UtilityDistrict, East Wenatchee, WA.

Hyatt K. D., C. Peters, P. Rankin, M. Stockwell, and C. Alex-ander. 2008b. The sockeye sub-model of the OkanaganFish/Water Management (OKFWM) Tool. In The Okana-gan/Fish Water Management (OKFWM) Tool: Record ofdesign (v. 2.4.000), ed. C. A. D.Alexander and K. D. Hyatt,97–135. 2013. Report prepared for Fisheries and OceansCanada, Vancouver, BC and Douglas County Public UtilityDistrict, East Wenatchee, WA.

Hyatt, K. D., E. Fast, M. Flynn, D. Machin, S. Matthews, andB. Symonds. 2001. Water management tools to increaseproduction of Okanagan sockeye salmon. Fish-Water Man-agement Proposal (21 June 2001) submitted to DouglasCounty Public Utility Division, Wenatchee, WA. CanadianOkanagan Basin Technical Working Group Secretariat.

Metcalfe, R. A., C. Chang, and V. Smakhtin. 2005. Tools tosupport implementation of environmentally sustainable flowregimes at Ontario’s waterpower facilities. Canadian WaterResources Journal 30(2): 97–110.

Richter, B. D., R. Matthews, D. L. Harrison, and R. Wigington.2003. Ecologically sustainable water management: Manag-ing river flows for ecological integrity. Ecological Applica-tions 13(1): 206–224.

Segrera, S., R. Ponce-Hernandez, and J. Arcia. 2003. Evolutionof decision support system architectures: Applications forland planning and management in Cuba. Journal of Com-puter Software and Technology 3(1): 40–46.

Senese, D. 2010. Amenity resources and rural change in theOkanagan Valley of British Columbia. In The rural-urbanfringe in Canada: Conflict and controversy, ed. K. Beasley,158–175. Brandon, MB: Brandon University, Rural Devel-opment Institute.

Stalnaker, C. B., L. Lamb, J. Henriksen, K. Bovee, and J. Bar-tholow. 1995. The instream incremental flow methodology:A primer for IFIM. National Biological Service, BiologicalReport No: United States Department of the Interior. 29.

Stockwell, M. M., and K. D. Hyatt. 2003. A summary of Oka-nagan sockeye salmon (Oncorhynchus nerka) escapement

survey observations by date and river segment from 1947to 2001. Canadian Data Report of Fisheries and AquaticSciences 1106.

Summit Environmental Consultants Ltd. 2002. Observationsand modeling of bed movement and potential redd scour inOkanagan River. Project report 635–13.01 prepared for theOkanagan Nation Alliance Fisheries Department, Westbank:BC.

Summit Environmental Consultants Ltd. 2010. Okanagan watersupply and demand project: Phase 2 summary report.Report prepared for the Okanagan Basin Water Board, Kel-owna: BC.

Summit Environmental Consultants Ltd. and Polar Geosci-ences Ltd. 2009. Okanagan hydrology state of the basinreport: Phase 2 Okanagan supply and demand project.Report prepared for the Okanagan Basin Water Board,Kelowna: BC.

Symonds, B. J. 2000. Background and history of water manage-ment of Okanagan Lake and River. Water Management,British Columbia Ministry of Environment Lands and Parks,Penticton: BC. http://a100.gov.bc.ca/appsdata/acat/documents/r18253/BackgroundandHistoryOLRS_1274988150212_5894e9abef0f64269169f31973bbcd60b145936f70338bdf3efb731b1ac62955.pdf (accessed September, 2014).

Symonds, B., B. Guy, B., L. Uunila, and R. Jubb. 2008.The socio-economic water management rules submodelof the Okanagan Fish-and-Water Management Tool.In The Okanagan Fish-and-Water Management Tool(Ok-FWMT): Record of design (v. 2.4.000), ed. C. A. D.Alexander and K. D. Hyatt, 136–161. 2013. Report pre-pared for Fisheries and Oceans Canada, Vancouver, BCand Douglas County Public Utility District, EastWenatchee, WA.

Turon, C., J. Comas, J. Alemany, U. Cortés, and M. Poch.2007. Environmental decision support systems: A newapproach to support the operation and maintenance of hori-zontal subsurface flow constructed wetlands. EcologicalEngineering 30(4): 362–372.

Watkins, D. W., and D. C. McKinney. 1995. Recent develop-ments associated with decision support systems inwater resources. Review of Geophysics 33(Suppl. no. 2):941–948.

Zar, J. H. 1984. Biostatistical analysis. 2nd ed. EnglewoodCliffs, NJ: Prentice-Hall.

Zhang, K., A. Zargar, G. Achari, M. Islam, and R. Sadiq.2014. Application of decision support systems in watermanagement. Environmental Reviews 22: 189–205.


Dow

nloa

ded

by [

Kim

Hya

tt] a

t 09:

34 0

4 M

arch

201

5

http://a100.gov.bc.ca/appsdata/acat/documents/r18253/BackgroundandHistoryOLRS_1274988150212_5894e9abef0f64269169f31973bbcd60b145936f70338bdf3efb731b1ac62955.pdf




The Okanagan Fish-Water Management (OKFWM) Tool: Balancing Water Objectives in Real Time

Documents