Counting fish: Performative data, anglers’ knowledge-practices and environmental measurement

Geoforum 43 (2012) 1014–1023

Contents lists available at SciVerse ScienceDirect

Geoforum journal homepage: www.elsevier.com/locate/geoforum

Counting fish: Performative data, anglers’ knowledge-practices and environmental measurement Sally Eden Department of Geography, University of Hull, Cottingham Road, Hull HU6 7RX, United Kingdom

a r t i c l e

i n f o

Article history: Received 3 June 2011 Received in revised form 13 April 2012 Available online 24 July 2012 Keywords: Environmental management Environmental measurement Lay knowledge Angling Water resources Fish ecology

a b s t r a c t This paper examines how environmental resources are measured and quantified as objects of environmental science and management and how lay knowledge-producers participate in this process, alongside the state. Using a case study of recreational angling, I show how fish in English rivers and lakes are counted and anglers act as lay or amateur knowledge-producers in the state’s metrological knowledgepractices. As embodied measurement instruments, anglers create data about themselves (as ‘effort data’) and about fish (as ‘catch returns’). These data are combined with other forms of data produced by the Environment Agency in England and Wales and used for fisheries management, thus shaping water bodies and fish ecology. I show how, to support environmental measurement, the state manages not only the environment and fish, but also anglers as lay knowledge-producers, using both regulation and economic incentives; in response, anglers also use data reflexively and strategically. I therefore emphasise the heterogeneous co-productions of environmental measurement as amateur–professional, human–animal and organic–technological, and show how measuring and managing water ecologies also involves measuring and managing humans. Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction This paper is about how fish in English rivers and lakes are counted and how anglers perform as embodied instruments of environmental measurement. Measuring environmental resources is rarely easy, but when those resources are invisible to normal human perception by being underwater and highly mobile – as fish are – then measurement becomes even more difficult. To count fish, environmental managers try literally to enrol their human predators – recreational anglers – to submit records of their catches as part of the process of measuring and managing fish stocks. These metrological practices transform fish into numbers, creating new ‘‘calculable objects’’ (Barry and Slater, 2002, p. 181) for environmental management. The varied performances by leisure anglers of casting a line, catching fish, counting fish, identifying species and recording counts not only render fish as ‘environmental resources’ that can be measured but also make those resources materially, through enacting different representations of environmental realities (Law, 2008) that then shape management of water environments. Drawing on work in human geography and sociology, I demonstrate how environmental measurement depends upon performative data or what the Environment Agency (EA) of England and Wales calls ‘effort data’, that is, data about time spent on recreational angling. Such calculative strategies of counting, measuring E-mail address: [email protected] 0016-7185/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.geoforum.2012.05.004

and categorising, of turning things into numbers (Barnes and Hannah, 2001; Elden, 2007) shape how we understand and manage environmental resources, yet these strategies are often hidden. My example is also unusual because it relies on literally thousands of members of the public who, as a byproduct of their environmental recreation, produce data about the fish they encounter (or fail to encounter) and about the time spent doing so. They are also themselves monitored and managed as measurement instruments by the state in the shape of the EA. In the process not only environmental data are created, but also relationships between state agencies and environmental recreationists. I therefore focus on three questions: 1. How are amateurs enrolled and managed as recreational knowledge-producers by the state to co-produce environmental measurement? 2. How are measurements from amateur recreationists and professionals combined and what happens when they disagree? 3. How do these co-produced measurements from amateurs shape environmental management and the materialities of water bodies specifically? These questions emphasise the heterogeneous co-productions of environmental measurement as amateur–professional, human–animal and human–technological, and show how managing water also involves managing humans reciprocally as instruments of environmental measurement.

S. Eden / Geoforum 43 (2012) 1014–1023

2. Measuring environmental resources My three research questions speak to the importance of counting, calculation and measurement in understanding and rethinking spatial politics generally (Crampton and Elden, 2006). On the one hand, turning environmental resources into numbers that can be managed has been seen as a modernist endeavour that supports capitalist control of the environment. St. Martin (2005) considered how fisheries science and management were used to bring the unruly environmental practices of commercial fish harvesting firmly within capitalism, especially through quantification and privatisation, making fish into abstract ‘resources’ ready for exploitation. Similarly, Demeritt (2001, p. 455) showed how the statistical measurement of American forests in the early 20th century supported new state agencies and ‘professional’ forest experts in managing those forests as an environmental resource, by transforming ‘‘heterogeneous forest stands into an apparently calculable quantity available to new forms of precise disciplinary control and governmental power’’. The recent focus on ‘ecosystem services’ in the UK also shows how practices of environmental measurement are linked with capitalism, as conversion to monetary units is seen as the most influential way to calculate the ‘value’ of environmental resources and therefore argue for their protection (e.g. UK NEA, 2011). On the other hand, social scientists have criticised this modernist conceit of control and explored how managing environmental resources is made problematic by proliferating hybrids and the difficulties of counting mobile, diverse and poorly defined populations such as bison (Lulka, 2004) and fish (Bear and Eden, 2008; Mansfield, 2003). Rather than seeing nonhuman entities as resources for capitalism, such analyses are more interested in tracing the processes by which ‘environmental resources’ are defined and made as objects of science and management, often drawing on actor-network theory and paying attention particularly to hybrid assemblages and knowledge-practices, as well as the embodied performance of knowledge production. Examples include mapping forest-savanna boundaries (Latour, 1999), classifying grassland types (Waterton, 2003), counting bird populations in cities (Hinchliffe, 2008) and defining American catfish or farmed salmon (Mansfield, 2003). In diverse examples, therefore, the knowledgepractices that create and use data are shown to be performative and creative of new realities (Law, 2008). Such analyses emphasise not merely the uncertainty in categorising living beings, but also the practical problems of counting environmental entities that move differently in time and space from the way that humans do and of making those entities matter for environmental policy (e.g. Hinchliffe, 2008). This is particularly problematic in the ‘field sciences’, such as ecology, which typically ‘‘resist tidy solutions’’ to measurement problems (Kuklick and Kohler, 1996, pp. 1–2). This brings us to the first of my research questions: how are amateurs enrolled and managed by the state as recreational knowledge-producers and thus co-producers of environmental measurement? Field sciences have often included amateurs or lay-people in environmental knowledge production, because organisations responsible for measuring and managing environmental resources in the public interest (usually for the state) have insufficient resources to do so themselves. Enrolling enthusiastic amateurs as (paid or unpaid) volunteers can extend the amount and scope of environmental fieldwork that conservation and environmental science can do, such as amateur naturalists collecting biodiversity data for the state (Ellis and Waterton, 2004, 2005; Meyer, 2010). This is also useful for non-governmental organisations with large memberships but small budgets. For example, the UK’s Royal Society for the Protection of Birds runs Garden Birdwatch, with up

1015

to 280,000 people participating by reporting bird counts, and the British Trust for Ornithology (BTO) has for decades relied on volunteers to collect data on British bird populations (Greenwood, 2003). In the USA, the Cornell Lab for Ornithology (CLO) and the National Audubon Society run the Great Backyard Bird Count, with over 80,000 counts submitted, as part of their ‘citizen science’ programme (Bonney et al., 2009). Such lay participants often need to be managed and nurtured. The BTO aim to give their volunteers ‘‘ownership of the work from the start’’ and ‘‘value and cherish them not just as fieldworkers but as a network of well-informed people who help take messages out to the wider community’’ (Greenwood, 2003, p. 228), including supporting training programmes for ringing birds (http://www. bto.org/volunteer-surveys/ringing/about/faqs). Nurturing may involve professional organisations or state agencies showing volunteers, such as anglers or birdwatchers, that the data they report are valued and useful to scientists and policymakers (Bell et al., 2008, p. 3450). Bonney et al. (2009, p. 981) reported that CLO’s ‘eBird’ website was improved to help people ‘‘to track their own observations and to explore how their reports compare with others’’ and the numbers submitting data ‘‘nearly tripled’’ afterwards, emphasising the usefulness of two-way communication in amateur–professional relationships. However, researchers tend to feel that amateurs should help out for what Ellis and Waterton (2005, p. 685) call ‘‘the wider public good, in the form of scientific knowledge’’ and for their own satisfaction, rather than for monetary recompense (Lawrence, 2006). Working for their own satisfaction is more problematic for fishers counting fish for the state. Commercial fishers have an economic incentive to provide data to the Canadian ‘Sentinel’ programme, because this should ensure that fishery management is based on good estimates, unlike the over-estimates of stock pre-1989 that led to over-exploitation and fishing bans (Finlayson, 1994). In England and Wales, the EA has not only a statutory duty to manage fish stocks, but also an economic incentive to increase angler numbers because (unlike many other outdoor recreationists) freshwater anglers must buy a rod licence annually from the EA. This is an important source of revenue for the EA but also makes anglers feel that they are owed a good service in return. Hence, EA staff (Aprahamian et al., 2010) note that ‘‘fisheries management is as much about people and geography as it is about fish stocks and ecosystems’’, because falling angler numbers do not reflect changing fish numbers (and thus the possibility/satisfaction of catching) but changing social issues (q.v. Eden and Barratt, 2010). ‘‘Promotional activity is extremely worthwhile’’ as a consequence, say the EA authors, because, unlike in France and the USA, ‘‘an investment of approximately 2% of licence income is preventing the decline in angling seen elsewhere and is delivering an average 6% increase in sales’’ (Aprahamian et al., 2010, p. 103). The economic aspect of the angler-state relationship therefore also shapes its knowledge-producing aspect and I return to this point later. My second research question asks how measurements from amateur recreationists and professionals are combined and what happens when they disagree. The literature shows how environmental scientists combine diverse data sets by translating living beings into counts, maps and graphs and classifying spatial and temporal variability. Latour (1999, p. 46) noted how forest scientists standardised recordings of samples in logbooks, to ensure later comparability; materialities (soil, plants) were transformed into numbers and codes, into writings and mappings, their differences gradually eroding so that they could travel through space and time. Standardisation classifies and transforms heterogeneous individuals from idiographic information into reductionist, aggregated abstractions, such as herds (Lulka, 2004), forests (Demeritt, 2001) or species.

1016

S. Eden / Geoforum 43 (2012) 1014–1023

But standardisation must be shared by different groups to work, often relying upon diverse ‘boundary objects’ (Star and Griesemer, 1989) that structure data collection and enable data to be readily shared by different user groups, especially amateurs and professionals. Boundary objects may be reference materials that define types (Waterton, 2003) or devices for recording counts on paper (e.g. tables or maps) or on computers (e.g. spreadsheets). However, boundary objects and other devices for standardising and sharing knowledge may be used differently by amateurs and professionals. Standardisation is therefore not one process but many, and may work or fail in different contexts. Moreover, authors dispute the value of amateur and professional knowledges. Bell et al. (2008) claimed that sometimes the expertise (knowledge) of ‘amateur’ naturalists is greater than that of ‘professional’ scientists, especially in-field recognition, whereas Danielsen et al. (2005) argued that local (lay) assistants in developing countries are less accurate than professionals.1 Greenwood (2003, p. 227) suggested a division of labour: ‘‘By using amateur birdwatchers, with field skills honed through long practical experience, we secure the validity of the basic observations; by using professional organisers, we ensure good survey design, proper collation of the data, and sound analysis.’’ In practice, any group of data-producing amateurs will include diverse abilities and levels of commitment. Identifying fish species, although simpler than other forms of classification like grasslands (Waterton, 2002) or mosses (Ellis and Waterton, 2005), can still be problematic even for experienced anglers, because defining separate species reflects not only fish differences but human perception, classification and generalised ‘‘information structures’’ (Bowker, 2000). The idea of a ‘species’ is a human construct and classifying nonhumans by species depends upon humans’ variable ‘perceptual skills’ (Ellis and Waterton, 2004). For example, sea trout spend most of their lives in the sea but return to rivers to reproduce, as salmon do; sea trout are therefore frequently confused with salmon despite being a different species (but the same species as brown trout which spend all their lives in rivers and lakes). Individuality and heterogeneity between animals and between anglers work against straightforward classification: not all fish from one species look alike to all anglers and fish from different species may look alike at different ages. The problem is that ‘‘the embodied present thing also often needs to be representative of other things in order to gain conservation value’’ (Hinchliffe, 2008, p. 89, emphasis in original) and ‘representative’ is often judged against ‘type specimens’ and their descriptions (Bowker, 2000, p. 650), perhaps in a museum or ‘classic’ paper. By comparison, the living entity in the field may fail to measure up: it may not look ‘right’, it may not behave ‘normally’ for its species, it may be in the ‘wrong’ place at the ‘wrong’ time. This unpredictability is indeed part of the fun that anglers experience – during ethnographic fieldwork, I have seen anglers choosing what they thought to be the ‘right’ bait to attract chub, but catching trout and pike instead (and still enjoying themselves). Sometimes, lay knowledge production and shared boundary objects may seem to blur the amateur–professional boundary (Star and Griesemer, 1989; Kuklick and Kohler, 1996, pp. 4–5), yet it persists spatially: the ‘amateurs’ are usually out in the field, whether this is defined as their garden or open countryside, and not in the centres of calculation (Latour, 1987), whether these spaces are laboratories or state offices. ‘Professionals’ may also go out into the field, perhaps to train the ‘amateurs’, but they come back to their laboratory, office and centralised database to perform the 1

This negative view may be an ethnocentric effect or a scientistic effect.

large-scale data accumulation, validation and presentation practices necessary for policy. Amateur naturalists enrolled by English Nature and the Natural History Museum to gather biodiversity data were required to submit their biological specimens and/or written records to professionals or ‘experts’ for verification and validation (Ellis and Waterton, 2005, p. 686). Checking is therefore part of the lay-expert relationship, although Ellis and Waterton suggest that the hope of the ‘expert’ in this case is that the ‘lay’ counter will learn and become more ‘expert’ through the process, making verification supportive, constructive and somewhat reciprocal, rather than merely critical. Similarly, a lot of fisheries data comes from commercial fishers. As I noted above, fish are particularly difficult environmental resources to measure because they are underwater (and thus invisible to normal human perception) and also highly mobile, so scientists lack good data about fish stocks, especially in the sea, and have often turned to both commercial and recreational fishers as cheap, quick and reliable sources of data. The fisheries literature uses both quantitative data from catch returns (by both net and rod) and qualitative data from interviews with fishers. For example, Murray et al. (2008, p. 589) refer to ‘‘fishers’ ecological knowledge’’ gathered from interviews and focus groups as FEK but, after the authors finished ‘‘gathering, transcribing, distilling and verifying’’ and thus improving it, they refer to it as FEK. The asterisk is important because it distinguishes scientists’ from fishers’ knowledge while valuing both and combining them as ‘‘a potentially valuable complement to existing tagging studies by providing a more complex picture of movements and stock structure at the local scale’’ (Murray et al., 2008, p. 593). Indeed, early fisheries science relied on information from local fishers, although Murray et al. (2008) suggest that this reliance has reduced as fisheries science has developed techniques to monitor fish directly, such as tagging programmes in the Canadian Atlantic. By tapping into fishers’ knowledge (an endeavour delightfully named ‘ethnoichthyology’), the fisheries literature has sought to exploit and value even anecdotal, qualitative data for practical management. Much of this work focuses upon commercial fishers, rather than recreational anglers, but Aprahamian et al. (2010) is an exception. There is also little sociological examination of how environmental resources are measured and diverse data combined, except for Finlayson’s detailed ‘‘forensic sociology’’ of the uncertainty and disagreements involved in northern cod estimates. Finlayson (1994, p. 102) found that state scientists tended to accept offshore commercial fishers’ knowledge but ignore that from inshore fishers, because the latter was seen as too anecdotal, qualitative and messy to collect from ‘‘a hopelessly heterogeneous muddle’’ of boats, emphasising the problems in combining heterogeneous data and collectors. Given these problems of partiality and heterogeneity, how do amateur knowledges shape environmental management and water materialities? Finlayson showed how estimates of cod stocks were used directly by the Canadian government to set catch quotas and how, when those estimates were exposed as vastly optimistic, the government suspended all fishing in an attempt (belatedly) to allow populations to recover. Implicitly, he suggested that if state scientists had listened to the small inshore fishers earlier, cod stocks would have been better protected. As a consequence of the cod collapse, the Canadian ‘‘Sentinel’’ programme was set-up by the state to formalise data collection from commercial sea fishers (Murray et al., 2008; http://slgo.ca/en/sentinel/context.html). This is a dramatic example of how defining and measuring environmental resources also makes those resources materially, because ‘‘realities (including objects and subjects) and representations of those realities are being enacted or performed simultaneously’’ (Law, 2008, p. 635, emphasis in original). Hence, the

S. Eden / Geoforum 43 (2012) 1014–1023

counts of fish also become fish as measurements are used to enact environmental management and thus co-produce environmental realities. In the case of anglers, their counts render fish as ‘environmental resources’ that can be brought within processes of science and policy, especially as aggregated data, to shape aquatic ecologies more widely. This compares with data collected by BTO volunteers about British farmland birds that has been used to indicate declines in species and thus biodiversity, but also to measure sustainable development generally (Greenwood, 2003).2 It is clear, therefore, that amateur metrologies do matter materially, as I shall show below. Together, varied counts, representations and standardisations across time and space transform things in ‘the field’ into numbers, words and reports (Latour, 1999) – in this case, they move fish from the water of the river to the paper of reports and policy, so that data are performative and creative of new realities (Bowker, 2000; Law, 2008). Latour (1999, p. 79) argues that the reversibility of such creativity anchors each stage and makes it more real as a constructed world. But reversibility is not so simple here – performances of classifying and counting can be less ordered and more unpredictable than might be expected (Waterton, 2002, p. 197), thus skewing the representations produced (Bowker, 2000, p. 659), and I return to this point later. 3. Measuring environmental resources through angling I now want to apply these three questions to recreational angling, which contrasts with the emphasis on commercial harvesting and other forms of ‘amateur science’ in the literature. I draw on a larger project studying anglers who regularly fished rivers in northeast England, for which two focus groups were recruited through local angling clubs: one of anglers who targetted coarse fish like chub, barbel, roach, dace and pike, and one of anglers who targetted game fish like salmon and trout. The difference between coarse and game angling is primarily about different target species or locations, but extends to differences in technology and practices (game species mainly targetted by fly fishing and coarse species by using bait) and class (fly fishing is often portrayed as more expensive and elitist). In practice, many anglers participate in both types over the years. After the focus groups, some participants were interviewed individually and more were recruited through club matches, snowballing and a fishing website. Sixty semi-structured interviews were completed in 2006–2008, plus extensive sessions of participant observation by rivers and lakes with anglers aged from 17 to 83 and with seven EA staff in the northeast of England (for methodological detail, see Eden and Bear, 2011a,b). We have elsewhere (Eden and Bear, 2011a,b) analysed the embodied practices of these anglers to demonstrate how they read the river and made sense of environmental change through recreational engagement. Others have also considered the bodily senses, technological materialities, spiritualities and masculinities involved in angling (Bull, 2009; Franklin, 2001; Snyder, 2007). In this paper, therefore, I do not focus on how anglers fish, but on how anglers count and how their counts are managed, which has not been dealt with elsewhere as a social practice. I include some qualitative data from interviews and participant observations, but because this paper differs from the original larger project, I mainly use EA documentation, phone interviews and email correspondence with EA staff in 2008–2011. Where interviews or personal emails are quoted, participants are given pseudonyms.

2 Defra have used BTO and other bird counts as part of its sustainable development indicators, although the precise usage changed over the years and the online information about this is now archived.

1017

I should highlight that game species like salmon and sea trout are counted differently from and potentially more thoroughly than coarse species like barbel, roach, chub and pike.3 The ‘fisheries statistics’ websites and reports produced by the EA deal solely with game fish, because these are more valued (and more valuable) than coarse fish. Salmon in particular is a charismatic species in a group (freshwater fish) that has very few charismatic species, so salmon is frequently used as a flagship species for river clean-up and rehabilitation, as well as attracting tourists to fly-fishing holidays. Differing charisma or appeal affects not only the conservation status of a species, but also its data status in terms of the glamour and funding attached to gathering knowledge about it (Bowker, 2000; Lorimer, 2006), which in turn will affect how much we know about and how we manage it. Knowledges about nonhumans are therefore discursive and material products of human effort and interaction with those beings. 3.1. Gathering measurements I now turn to my example of anglers and first consider how recreational anglers are enrolled and managed by the state as co-producers of environmental measurement. Anybody over 12 years of age fishing inland waters in England and Wales must buy a rod licence (for either salmon and migratory sea trout, or coarse fish) from the EA, which records their name and address. Anglers are thus enrolled as data collectors because, under the Water Resources Act 1991, all holders of rod licences to catch salmon and migratory sea trout ‘‘are legally required to submit a full and accurate catch return to the Agency’’ by the year end (EA, 2005, p. 4), reporting how many fish they caught and when. To manage data collection, the EA sends licence holders a small printed logbook. This serves as a boundary object (Star and Griesemer, 1989) by standardising how counts are recorded by anglers, e.g. in terms of units, content, layout and temporal resolution, and making these useful to the EA, thus bridging the two groups and enabling data sharing. The EA asks each angler to write each fish caught into their logbook ‘‘as soon as reasonably practicable and in any case before midnight on the day the event occurred’’ and they should also record ‘‘the time spent fishing’’ (what is later referred to as ‘effort data’) and ‘‘the individual weight of any salmon or migratory trout caught even if the fish is released or, where no fish were caught, a statement to that effect’’ (EA, 2009, p. 3). Since 1994, the EA has sent reminders to all licence holders at the end of fishing seasons (these vary by region) and claims to have reduced ‘‘under-reporting’’ to less than 10% (Table 1; EA, 2005, p. 4; EA/CEFAS, 2009, p. 49). Management also enrols fishing companies by requiring commercial licence holders to report their catch. But anglers’ data are more important: in 2008, the total catch of salmon in England and Wales (including fish caught and subsequently released by anglers and fish caught and killed by commercial fishing operations) was estimated at 112.4 tonnes or 30,800 fish, of which about 22,000 fish were reported by recreational anglers (EA/CEFAS, 2009): 71% of the national salmon count was therefore produced by anglers and dependent upon the quality and accuracy of their reporting. Coarse fish can be more problematic to count than game fish, because they tend to be caught by anglers in bulk, not as large specimens (except for trophy fish, such as barbel and pike), so are rarely individually counted like salmon. Instead of issuing individual logbooks, the EA enrols anglers into data collection for coarse fish collectively, by asking angling club secretaries to collate 3 This paper excludes important target fish like carp that live in ponds because, although ponds are connected to rivers via flooding, cross-infection or release, they do not greatly influence the wider management of rivers and the sea.

1018

S. Eden / Geoforum 43 (2012) 1014–1023

Table 1 Rod licences and catch returns for salmon and migratory trout in England. Source: EA/CEFAS, 2009.

1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 a

Annual rod licences issued

% Annual licence holders submitting catch returns

Short term (1day or 8-day) rod licences issued

% Short term licence holders submitting catch returns

21,146 21,161 19,223 14,916a 19,368 21,253 22,138 23,870 22,146 23,116

78 76 71 83 94 86 83 88 88 88

11,364 10,709 10,916 9434 10,039 8683 10,628 10,170 9460 9065

51 53 53 61 60 44 48 55 54 54

Table 2 Variability in match returns (for coarse fishing) received by EA from two large angling clubs in northern England. Source: EA (pers. comm., 2008), Frear (2008) and Lee (2009).

1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004d 2005 2006 2007 2008 2009

Low due to Foot and Mouth Disease (see page 1020 for explanation).

and submit data about the number of anglers fishing in a match, the number of anglers ‘weighing in’ their netted catch at the end, the total weight of fish and the species of fish caught. The availability of these data recorders is, however, strongly spatialised. Angling clubs in England and Wales rent or own fishing rights to fragments of rivers, sometimes as short as 500 m and often only on one bank, producing patchy geographies of access that also shift as rights are sold or transferred between seasons. Patchiness of club space produces patchiness of reporting by match secretaries – some clubs may hold no matches on particular stretches in a given year, some may hold many and sometimes the two may even hold rights to opposing banks of the same stretch. Amateur knowledge-producers may therefore not record in particular places: coarse fishing returns were routinely not completed for the River Aire in northeast England until 2008 due to pollution, leaving it out of counts completely. Ellis and Waterton (2005) contrast the cartographic imagination of the (state’s) conservation institutions that harvest the data centrally against the individual, ethnographic imagination of the amateur naturalists, characterised by various in/exclusions, often self-imposed, so that some areas may be under-reported and some over-reported. Fishing may also concentrate close to good car parking and (perceived) better quality spots as anglers become more selective about where to fish during matches (Lee, 2009, p. 20), emphasising the importance of anglers’ ‘effort data’ as well as fish availability. And anglers may resist enrolment in the EA’s metrological practices in other ways. The literature suggests that amateur measurers are less likely to record negative data (absence of species or conditions) than positive data, thus skewing results through ‘‘a tendency to overreport certain species and to underreport others,’’ especially common species (Bonney et al., 2009, p. 980). In the case of angling, anglers who are catching little or nothing in a match may leave early without bothering to ‘weigh in’ their net and thus their catch (if any) will be excluded from the data submitted by match secretaries; up to 60% of anglers in matches in the Dales area of northern England probably do not bother to ‘weigh in’ (Frear, 2008). Absence of animals, like presence, is also produced by metrological knowledge-practices within the professional-amateur network. Hinchliffe (2008, p. 91) noted that recording black redstarts in urban areas only ‘‘provides glimpses and fleeting presences. Absence from records does not necessarily mean ‘not present’ and recorded presence may be more or less significant depending on the behaviours of the bird.’’ In the case of anglers, the number of humans made present or absent by ‘weighing in’ will vary. For example, if prize money is distributed across several sections, more anglers will bother to ‘weigh in’, because even if their own section is fishing poorly, they may still be doing better

Leeds and district ASA % match returns

York and district AA % match returns

48 58 Not reported 28 0** 14 15 13 63a 87 62 62 12b 44 50 58 83 98 97

28 45 Not reported 41 47 38 23 14 17 24 Not reported 22 64c 51 85 87 87 87

Total number of returns in EA dales area

435 461 502 507 493

a

Leeds introduced compulsory returns at EA’s request. Affected by Foot and Mouth Disease outbreak – this is the combined total. York and district began encouraging club returns. d EA bonus scheme introduced. ** See page 1021. b c

than their neighbours in that poorly performing section (Lee, 2009). So the club’s arrangements in time and space enact reporting and thus affect the fish present in official counts. But even if all the anglers participating in a club match ‘weigh in’, the EA still depends upon the match secretary submitting the returns (maybe 20 or more) each year. Returns were declining in the 1990s, so the EA targetted large angling clubs (with memberships in the thousands) to more efficiently recruit counts by offering a ‘bonus’ payment. Table 2 shows how successful this was. Each return here represents 1 day’s catch from several (or many) anglers, rather than a full year for an individual game angler, so data collection differs by time and space, as well as by species, in that the rivers and lakes that are counted are those that are more popular with anglers and have more active clubs. Measuring and managing humans is thus part of measuring and managing nonhumans, producing varied geographies of environmental measurement as ‘effort data’ is factored into fish counts. For coarse fisheries, angler counts are also checked by the EA, who apply ‘Analytical Quality Controls’ to check whether data submitted by match secretaries are ‘‘valid’’, that is, whether counts are higher/lower than ‘‘normal’’, where ‘‘normal’’ means expected by EA staff given long term trends: ‘‘we know because we have such a robust data set within a 10-year mean’’ (Phone interview with Grant, EA). For example, if the EA receive a match return that seems (intuitively) very large, they would check it by phoning the match secretary ‘‘who might say, ‘Oh, no, that was a typo, it was 20’’’ not 200. EA staff also ‘‘train’’ the match secretaries in person annually and keep in touch: ‘‘it’s a comms thing. . . we have to be very proactive’’ in educating anglers and encouraging them to participate. Anglers pay the EA for a rod licence, but the EA also pays angling clubs to submit data, with an incentive of £1 per match return plus a ‘bonus’ of £200 a year to a club if they return at least 85% of matches – most clubs in the northeast now do. Similarly, the EA proposed to enter anglers who returned a one-off River Tyne

S. Eden / Geoforum 43 (2012) 1014–1023

logbook in 2009–2010 into a prize draw (EA http://www.environment-agency.gov.uk/research/library/publications/40717.aspx) as an incentive. So, the state manages humans as measurement instruments in order to manage nonhumans. And enrolment differs: like the two categories of fish (game and coarse), there are two categories of anglers using two categories of recording devices (game logbooks filled in by the individual and coarse match returns filled in collectively by the secretary on behalf of the club). Enrolment is encouraged both through regulation and economic incentives respectively, as well as communication. 3.2. Combining measurements Following enrolment, I now consider how measurements from amateurs/recreationists and professionals are combined and what happens when they disagree. The logbooks submitted by individual anglers are compiled centrally by the EA and the Centre for Environment, Fisheries and Aquaculture Science (CEFAS), then published as national fishery statistics (EA, 2005; EA/CEFAS, 2009). In this way state organisations ‘harvest’ data from ‘donating communities’ (Ellis and Waterton, 2004, 2005). Aggregated data are then used to evaluate environmental resources: for example, the fishery of the (small) River Esk in Yorkshire was measured through 186 returns by anglers in 2005, representing 1351 days fished and 111 salmon and 376 sea trout caught. Anglers’ raw counts are then recalculated, echoing Murray et al.’s (2008) FEK above, by weighting fish counts by angler counts. EA/CEFAS use angler ‘effort data’ in terms of hours to work out catch per unit effort (CPUE) – a standard measure in the fisheries literature internationally whereby the amount of fish caught is divided by the amount of time spent fishing, to control for the effects of changing angling practices. In addition, national ‘exploitation rates’ for salmon are calculated from ‘standard fishing units’ (i.e. rod licences issued) and weighted by anglers’ ‘‘catching power’’ or productivity, using CPUE but also assuming ‘‘a natural mortality rate of 3% per month’’ of salmon in the sea. ‘‘Whilst this model is acknowledged as containing a number of uncertainties, it provides our best interpretation of available information on salmon stocks at a national level.’’ (EA/CEFAS, 2009, p. 84) Catch returns from anglers and commercial companies are not the only source of fish counts. The EA also gathers its own primary data using remote devices and manual traps, what it often calls ‘‘independent measures’’.4 For example, in northeast England, the EA runs automatic fish counters on the three major rivers: the Tyne, the Wear and the Tees. Each counter has three electrodes in the stream bed that measure resistivity and send signals which can be interpreted as fish counts. The Tyne also has an underwater camera to confirm fish counts visually (EA http://www.environment-agency.gov.uk/research/library/publications/40723.aspx). Such monitoring offers correctives for data produced (in much larger amounts) by fishers, in the EA’s view: ‘‘Care is required in trying to draw general conclusions about current stock status from catches alone. The actual relationship between catch and stock abundance depends upon exploitation rates (i.e. the proportion of the salmon population taken in the catch – both retained fish and those released). This can be estimated where there is a fishery-independent measure of the salmon run, such as that obtained from fish counters.’’ (EA/CEFAS, 2009, p. 12, emphasis added)

4 The EA also measure river flow by hydrological and other data that are not covered in this paper.

1019

The EA also captures fish using traps or electricity to stun fish (‘electric fishing’), allowing staff to examine fish by hand to identify their species, age and health - for example, salmon are checked for red vent syndrome. The EA’s National Fishing Monitoring Programme stipulates a sampling timetable for electric fishing with two spatial components: ‘key’ or ‘temporal’ sites are surveyed annually to monitor change and ‘‘to provide a long term data set’’, whereas ‘spatial’ sites are surveyed on a rolling 6-year programme to monitor changes in the spatial distribution of fish (Phone interview with Grant and email correspondence with Grant and Anna, EA). The intensity of monitoring varies by size: the Ouse has seven key sites and four spatial sites, but the larger Swale has about ten and 20 respectively. In Scotland, fish trapped on the River Dee before 1 June in 2007 and 2008 were marked with coloured tags and then released; anglers were asked, if they re-caught fish with tags, to remove the top part of the tag and return it to the EA for a reward, but leave the lower part of the tag on the fish so that it could be identified again later (EA/CEFAS, 2009, p. 38). Here, knowledge-practices used to measure populations (as counts) differ from those used to measure individuals (as handled bodies). Approaches to collecting and managing data vary within the EA, with rules ‘‘often not formally written down’’ (Phone interview with Ed, EA), hindering retracing the precise steps of knowledge production. But it is clear that measurements from different types of collectors (humans and nonhumans) are combined: counts of population from anglers’ returns combined with samples of individual fish bodies from trap data (‘‘monthly age-weight keys’’) were extrapolated to produce a statistical profile of fish stocks by age ‘‘scaled up to the total catch (rods and nets combined) on a pro-rata basis’’ (EA/CEFAS, 2009, p. 83). I asked EA staff what happens when data from catch returns disagree with data from ‘independent measures’. They felt that this rarely happened, not least because individual variations or errors cancel each other out when thousands of angler-days are involved, producing ‘‘a strong correlation between counts and captures’’ (Phone interview with Ed, EA). However, electric fishing catches younger fish than anglers catch yet also catches fewer of some species, like bream, making the EA ‘‘reliant on catch data’’ for them (Phone interview with Grant, EA). On the rare occasions where EA staff see data mismatch, they say that angler counts tend to be lower than ‘independent measures’ (rather than vice versa), because something has prevented the anglers catching fish (bad weather, poor technique, etc.). A corollary is that anglers might have a more negative view of stocks than the EA, increasing their criticism of how the EA is managing fisheries. Clearly, anglers’ data are valuable for the EA and usually felt to be in line with the EA’s own data and long-term expectations, although one EA officer did comment that, in a pinch, he would on balance trust data from electronic counters more than data from anglers. Yet the EA’s ‘independent measures’ are also sometimes flawed. The EA’s fish counter on the Yorkshire Esk worked from 1998 to 2003, but was then washed away in floods and had not been re-installed by 2010; another on the Coquet was discontinued because of cost and doubt about data reliability (Phone interview with Ed, EA). As well as the problems of controlling anglers’ data, the EA cannot control the weather or technological failure. There are also far fewer traps and counters than there are human fishers providing returns, so the spatial coverage of ‘independent measures’ is even more problematic. And data from ‘independent measures’ may still be uncertain. Signals from automatic fish counters are classified by a computer algorithm into ‘ups’ (fish migrating upstream), ‘downs’ (fish migrating downstream) and ‘events’ (non-fish). Where the algorithm cannot resolve the signal automatically, EA staff check visually for a

1020

S. Eden / Geoforum 43 (2012) 1014–1023

distinctive ‘sine wave’ (http://www.environment-agency.gov.uk/ static/documents/Research/counter__video_897529.pdf) as ‘‘a nice clear sign is definitely a fish’’ (Phone interview with Ed, EA), suggesting that errors arise but also that people are more accurate (but more expensive and slower) than automatic devices. And counts are also temporally differentiated, showing strong seasonal or yearly changes. In 2001, Foot and Mouth Disease prevented access to much of the English countryside, many angling matches were cancelled and catch returns therefore not produced (Table 2). In 2007, many matches in northern England were cancelled due to flooding, reducing data coverage but also discouraging anglers from participating (and reducing ‘effort data’) even when matches did occur, because high water levels can be dangerous or difficult to fish. Changing environmental conditions can also favour some fish species being counted more than others: high water and suspended sediment, for example, are seen as conducive to catching barbel rather than chub, influencing how anglers perform their fishing and (more arguably) how fish respond: ‘‘Elevated river flows and high turbidity reported on most match return cards for this venue will have undoubtedly favoured barbel fishing at the expense of other species.’’ (Frear, 2008, p. 8) However, the EA seems wary of claiming too much for these fish counts – they serve more as relative measures of change between rivers and times (especially times of different management regimes), rather than as absolute measures of species stocks. Hence, ‘‘data is only valuable when something goes wrong. . . everything is risk-based’’ (Phone interview with Grant, EA), so the real value of data is in showing long term trends as a benchmark for identifying problems or damage, rather than one-off, seasonal changes. For example, increased perch catches reported by anglers on the Aire in 2009 were used by the EA as proxies for water quality improvements following management changes (Lee, 2009, p. 20). 3.3. Using the data I now consider how these combined measurements shape environmental management and water materialities. Data on game fishing inform EA/CEFAS reports which are sent to the International Council for the Exploration of the Seas (ICES) which in turn advises the North Atlantic Salmon Conservation Organisation (NASCO) which manages salmon stocks and negotiates quotas for salmon fisheries in the North Atlantic. Anglers’ counts of fish in English and Welsh rivers thus shape international decisions about how to manage marine stocks. Data on coarse fishing are used by the EA to evaluate rivers using a Fisheries Classification system for England and Wales. Classes A–D are ‘‘a representation of its fisheries’ status and quality’’ (Frear, 2008, p. 4), calculated as average Catch Per Unit Effort (CPUE) in grams of fish caught per angler per hour. Here the ecological reality of water quality is inscribed and classified through partial recordings of anglers’ ‘effort data’ – data that quantifies the anglers and their fishing behaviour, rather than quantifying fish populations directly, producing human–nonhuman environmental measurements. For example, the River Swale was classified based on data from only 11 days’ angling in 2007 and the River Ure from 15 days, all from the same (small) angling club and (likely) the same few anglers. And diverse types of data are combined in a ‘‘State of the Nation’’ fisheries report for salmonids and coarse fish, roughly every 10 years. This ‘‘management tool’’ (Phone interview with Grant, EA) is thus a hybrid co-production, being human– nonhuman, amateur–professional and organic–technological. Data specifically from electric fishing is used identify and manage water

bodies that do not meet ‘good ecological status’ under the EU Water Framework Directive. The consequences of these measurements can be various temporal and spatial fixes: fishing may be banned at particular times, the number of commercial licences may be reduced, methods may be restricted to specified artificial flies and lures or anglers may be restricted to catching only two fish per day on some rivers. Data become reflexive metrologically – anglers’ data suggest that fish populations are under stress, thus influencing management, thus influencing (but not determining) how many fish anglers catch (and count) in future as populations change. Recreational predators thus become part of the process of protection. For example, counter data were used to assess the effects of constructing the Tyne Tunnel and the Tees barrage on fish migration upstream. When a problem was identified on the Tees, the EA asked the operator to change the barrage regime to improve water flow, to literally raise the water level as a direct hydrological (and an indirect ecological) data outcome. Similarly, counter and angler data are combined to monitor stocks in the Tyne and may prompt more water to be released from Kielder reservoir to raise water levels downstream. For coarse fish, angler data from the Nidd was used to assess the impact of Skipbridge weir on fish migration (see http:// www.environment-agency.gov.uk/hiflows/station.aspx?27062). In 1978–1980, anglers reported falling catches on the Nidd. The EA reviewed the long-term data which supported the anglers, so various measures were tried to raise the water level but data continued to show that the weir ‘‘didn’t work’’ (Phone interview with Grant, EA; http://www.environment-agency.gov.uk/business/sectors/37579.aspx). The EA then used a camera to record fish being unable to leap past the weir and showed the video to the regional fisheries committee. This proved more persuasive than graphs of falling numbers and the weir was removed in 1999; afterwards, catch data showed fish populations increasing to pre-weir levels. This story shows clearly the combination of angler data, EA data and a final explicitly visual (and qualitative) act of ecological politics that literally shaped a river.

4. Complexities, reciprocity and (ir)reversibility So far, I have shown how catch returns are turned into fish numbers and used in management, such that logbooks and ‘effort data’ embody ‘‘a piece of nature mediated by the effort of human observation’’ (Ellis and Waterton, 2005, p. 684), that is, fish mediated by humans. This has several wider implications. First, there is the reciprocal nature of the relationship between fishers and the state. Anglers pay the EA to be allowed to fish but the EA pay angling clubs to provide catch data and the EA also give the data back in the form of free reports, e.g. detailed fish counts at http://www.environment-agency.gov.uk/ research/library/publications/40723.aspx. There are other incentives too: in interview, Grant said that ‘‘the biggest carrot’’ for anglers to submit catch data is that the EA ‘‘give it back to them’’ when they need it to argue a case, such as when applying to Natural England for licences to shoot cormorants that they think are depleting fish populations, or after a pollution incident to show fish decline against long term trends or after tree planting. Like fish, anglers are both measured and managed, but through different knowledge-practices – through ‘effort data’ and CPUE, through regulation, bonus payments and reminders respectively. The EA seeks to enrol and nurture relationships with anglers, encouraging them ‘‘to have some ownership of the data’’ (Phone interview with Grant, EA). But the EA also uses the anglers’ data to demonstrate that changing fish populations are ‘‘natural’’, that is, not anybody’s fault, and thus to counter arguments from anglers

S. Eden / Geoforum 43 (2012) 1014–1023

that fisheries need ‘fixing’. Hence, counts from anglers are politicised. And because they know (or imagine) that the EA use their catch returns and ‘effort data’ in this way, anglers may fill logbooks in strategically, rather than accurately. Like tactical voting in marginal constituencies during political elections, anglers may use tactical recording of catches to de-emphasise success and highlight a fishery needing improvement. Hannah (2001) calls strategic in/exclusion in the US Census ‘statistical citizenship’, a political act on behalf of the self or other under-represented humans; here it is (at least partly) a ecopolitical act on behalf of non-represented nonhumans, i.e. fish. But reciprocity also relates to the potential for knowledge reversibility. Latour (1999, pp. 70–74) refers to a reversible, traceable chain of reference from matter to form and back again – metaphorically flowing both ‘upstream’ and ‘downstream’ – to emphasise that one must be able to ‘read back’ from an item where it came from. But this is clearly problematic where aggregation of fish begins at the very first count (the angler’s net or rod) and it is not possible for angling clubs, the EA or any interested party to work ‘upstream’ beyond that without additional resources (such as doing fieldwork themselves): some transformations of knowledge become ‘‘practically untraceable’’ (Waterton, 2002, p. 197). This irreversibility is compounded when databases are built across multiple institutions, countries and disciplines into large and multi-layered (discursive and material) constructions, as for international and migratory fish populations: ‘‘first because the infrastructure is performative; and second because the infrastructure is diffuse’’ (Bowker, 2000, p. 648). EA staff (Aprahamian et al., 2010) used logbook data to test the effect of limiting and banning fish catches on the Lune and the Usk respectively, combining different data types (anglers, automatic counters) in their results without annotation or explanation, blocking reversibility and retracing. The ‘data histories’ (Bowker, 2000, p. 675) needed to reverse-engineer the data later (Latour, 1999) are lost. So, we do not merely need good practices in terms of gathering knowledge, but also good practices in terms of sharing and archiving it, especially where the original practices become less visible (and their choices irreversible) as heterogeneous types of data, gathered through heterogeneous practices, places and people, are combined, aggregated and the early differences largely erased. But it is impossible to get deep enough into the data to identify how much calculation is literally dependent upon performance, upon the anglers’ bodily skills. Anglers choose baits, flies and line to suit their target species and this influences but does not determine (because fish retain agency and behave unpredictably) the fish caught and thus the fish reported. And because only fish that have been counted are valued and protected, this again emphasises ‘‘the performative aspect of the infrastructure’’ (Bowker, 2000, p. 659) and the ‘‘flickering associations’’ (Hinchliffe, 2008) between presence of species and measurement of species. Angler skills and effort vary across time. Some anglers may fish more or less often than they had before; some anglers may give up; new anglers may start; some anglers may fish better (catching more) or worse (catching less) than they did before because of a whole range of personal factors, including technological upgrades, changing use of bait, greater tolerance of poor weather and greater experience enhancing their skills. Standardisation becomes impossible though, because of the fun factor – anglers produce data as a by-product of their recreation and thus in practices that are less controlled, less codified (except in the methods of data recording) than in other ‘amateur science’ initiatives. This is why the EA/CEFAS (2009) refer to ‘effort data’ as well as ‘catch returns’ when reporting fish counts, acknowledging that counts of fish in anglers’ catch returns are also counts of the frequency, density and variability of anglers’ behaviour (rather than fish behaviour) across time and space. But even ‘effort data’ that measure anglers’

1021

input in terms of time cannot quantitatively factor in anglers’ skills and the effects of technologies. As well as angling behaviour, counting fish also depends upon anglers’ form-filling and writing behaviour which, like their angling, will vary across time and space. Again, this reflects a reciprocal relationship with the state in that many anglers see the EA as highly bureaucratic, which discourages them from complying with EA reporting requirements. For example, the reported number of ‘rod days’ that anglers fished for salmon or sea trout in England and Wales declined from nearly 300,000 in 1993 to around 200,000 in 1999 (EA, 2005, p. 7) and then stabilised for several years (except 2001, due to Foot and Mouth Disease). But it is a moot point as to whether anglers were fishing less in 2005 than in 1993 or merely reporting less. In a more explicitly politicised example, in 1993 anglers ‘‘rebelled’’ against a government proposal to raise a levy on fishing venues by lobbying government, but also by not providing any match returns to the agency that would be charged with collecting the levy (see Table 2); the proposal was later scrapped (Grant, EA, pers. comm. 2010). It could be argued that temporal variation (especially over short time periods) is even more pronounced in water environments than it is in land environments, such as the forest studied by Latour (1999) and the grassland by Waterton (2002). Water levels in rivers can rise metres in days or even hours after heavy rainfall, changing the size, speed and quality (e.g. suspended sediment) of the water and thus changing how fish and anglers behave. And unlike plants, fish are highly mobile, moving between stretches of river, between different rivers and between rivers and the sea over time, as well as spawning in different seasons. Capturing temporal variation demands complex calculative strategies, but remains incomplete. And unlike Latour’s (1999) forest scientists, anglers may fill in their logbooks ex situ, away from the river banks and comfortably at home and they may lose their logbook and be sent a replacement to fill in ‘‘to the best of the holder’s knowledge’’ (EA, 2009, p. 3), in both cases relying on hindsight and imperfect memories maybe weeks or months later – another form of temporal uncertainty.

5. Conclusions I have shown how angler’s ‘effort data’ (both fishing and formfilling) are enrolled into the calculative strategies of state agencies like the EA and in turn shape water resources. Unlike other cases of amateur data collection (Bonney et al., 2009; Ellis and Waterton, 2005; Greenwood, 2003), anglers’ ‘skill’ or ‘expertise’ is not necessarily in cognition or ecological identification (although this develops as a byproduct of their practice), nor is recording (on paper or on computers) their primary purpose. Rather, the embodied performance of catching fish is what matters for their recreation and participation - their inscriptive performance of filling in forms and submitting data (and thus knowledge production) is a by-product, rather than their main activity. Because of this, anglers can be more active in shaping (choosing) the conditions, spaces and times by which they produce and submit environmental knowledge, prompted not by the requirements of the state but by their own environmental imaginaries and recreational choices. This is why their ‘effort data’ are so explicitly factored into environmental measurement. Yet even their ‘effort data’ fail to capture much of their embodied performance and how this influences the catch data that result. Like other recording instruments, anglers produce data imperfectly and variably. Ecological assessment is ‘‘highly embodied, highly specific and localised to our own positioning in the field’’ (Waterton, 2003, p. 121) and such performances are also reflexively monitored by most anglers (some to the point of obsession)

1022

S. Eden / Geoforum 43 (2012) 1014–1023

because they want to improve. However, some anglers are more diligent and accurate than others, not only in terms of their embodied skills and ability to catch fish, but also in terms of identifying the fish that they do catch and filling in the forms thoroughly and precisely afterwards. Also, anglers know that their counts are used by the EA, so may adjust their counts strategically and reflexively, contextualised by their own views of what needs ‘fixing’ in the river ecologies that they encounter. By shaping the objects of management in ways that policy is more able to grasp, amateur knowledge-producers know that it is more likely that policy will continue to reproduce or maintain those objects – be they species, habitats or attributes – rather than others. ‘‘If we are only saving what we are counting, and if our counts are skewed in many different ways, then we are creating a new world in which those counts become more and more normalized’’ (Bowker, 2000, p. 675). Performativity of data collection is thus not only implicitly political in shaping environmental management, but also sometimes an explicitly political act. And the state’s own ‘independent measures’ are also subject to slippages and imperfections in time and space. Control and certainty are not the exclusive preserve of the professional or the state, as they are sought but rarely (if ever) achieved over environmental resources and their measurement. The environmental resources measured through simple counts of fish are thus heterogeneously produced by an array of professional and amateur, human and nonhuman, organic and technological, present and absent practices. I have also emphasised the reciprocal relationship between the state (in the form of the EA) and amateur data producers (in the form of recreational anglers). The EA tries to shape anglers as (better) recording subjects to subsidise environmental monitoring, but anglers also use their recording to shape state management, such as strategically reporting data or using it to argue for particular management practices. And anglers themselves are measured by ‘effort data’ and CPUE, being part of counts rather than merely objective instruments for counts. Ellis and Waterton (2004) suggested that amateur enrolment in environmental measurement is a new role for publics in environmental science and policy, through citizenship that uses amateurs not merely to get good environmental data (a ‘‘data-led’’ or ‘‘extractive’’ strategy), but to build public capacity and better state-public relationships through coproducing environmental knowledge that is fully contextualised within those relationships. I have shown here that not only does the state manage amateur anglers as recording instruments but that anglers are reflexively aware of this role and thus may strategically report and use data deliberately to support their own agendas of water management, an approach that is far from an ‘extractive’ strategy. Where environmental knowledge is co-produced through relationships – whether human–nonhuman in terms of catching fish, or human–human in terms of sending completed logbooks to the EA – these relationships imply a duty of care. Data are not produced without value and recognising amateurs’ efforts is important (Ellis and Waterton, 2004; Bell et al., 2008) for not only the flow of data, but also the quality of it. In the case of anglers in England and Wales, for example, the EA could do much more to communicate the value of data to anglers, to build more reciprocal, positive and long-lasting relationships with them, not by rewarding data-collectors only through economic payment but also by sharing knowledge and its value for decisionmaking more widely with its angling public (although with the EA’s budget shrinking, this is looking unlikely). More generally, measuring environmental resources and producing environmental knowledge should be seen not merely as exploitation (of both nonhumans and humans through labour), but as a moral relationship that ties together

different groups and ecological realities both in the present and in the future.

Acknowledgements This paper draws in part on work funded by the UK Research Councils’ Rural Economy and Landuse (RELU) programme, under ‘Angling in the Rural Environment’, award RES 227 25 0002. I would like to thank Environment Agency staff (especially Paul Frear) for helping me understand how the Environment Agency collects data and for comments on an earlier version of this paper. I would especially like to thank Chris Bear, who conducted some of the angler interviews referred to in this paper, provided useful comments on an earlier version and also was key to the success of the RELU project over the years. References Aprahamian, M.W., Hickley, P., Shields, B.A., Mawle, G.W., 2010. Examining changes in participation in recreational fisheries in England and Wales. Fisheries Management and Ecology 17, 93–105. Barnes, T.J., Hannah, M., 2001. The place of numbers: histories, geographies, and theories of quantification. Environment and Planning D: Society and Space 19, 379–383. Barry, A., Slater, D., 2002. Introduction: the technological economy. Economy and Society 31 (2), 175–193. Bear, C., Eden, S., 2008. Making space for fish: the regional, network and fluid spaces of fisheries certification. Social & Cultural Geography 9 (5), 487–504. Bell, S., Marzano, M., Cent, J., Kobierska, H., Podjed, D., Vandzinskaite, D., Reinert, H., Armaitiene, A., Grodzin´ska-Jurczak, M., Muršicˇ, R., 2008. What counts? Volunteers and their organisations in the recording and monitoring of biodiversity. Biodiversity and Conservation 17, 3443–3454. Bonney, R., Cooper, C.B., Dickinson, J., Kelling, S., Philips, T., Rosenberg, K.V., Shirk, J., 2009. Citizen science: a developing tool for expanding science knowledge and scientific literacy. BioScience 59, 977–984. Bowker, G.C., 2000. Biodiversity datadiversity. Social Studies of Science 30 (5), 643– 683. Bull, J., 2009. Watery masculinities: fly-fishing and the angling male in the South West of England. Gender, Place and Culture 16 (4), 445–465. Crampton, J.W., Elden, S., 2006. Space, politics, calculation: an introduction. Social & Cultural Geography 7 (5), 681–685. Danielsen, F., Burgess, N.D., Balmford, A., 2005. Monitoring matters: examining the potential of locally-based approaches. Biodiversity and Conservation 14, 2507– 2542. Demeritt, D., 2001. Scientific forest conservation and the statistical picturing of nature’s limits in the Progressive-era United States. Environment and Planning D: Society and Space 19, 431–459. Eden, S., Barratt, P., 2010. Outdoors versus indoors? Angling ponds, climbing walls and changing expectations of environmental leisure. Area 42 (4), 487–493. Eden, S., Bear, C., 2011a. Reading the river through ‘watercraft’: environmental engagement through knowledge and practice in freshwater angling. Cultural Geographies 18 (3), 297–314. Eden, S., Bear, C., 2011b. Models of equilibrium, natural agency and environmental change: lay ecologies in UK recreational angling. Transactions of the Institute of British Geographers 36, 393–407. Elden, S., 2007. Governmentality, calculation, territory. Environment and Planning D: Society and Space 25, 562–580. Ellis, R., Waterton, C., 2005. Caught between the cartographic and the ethnographic imagination: the whereabouts of amateurs, professionals, and nature in knowing biodiversity. Environment and Planning D: Society and Space 23, 673–693. Ellis, R., Waterton, C., 2004. Environmental citizenship in the making: the participation of volunteer naturalists in UK biological recording and biodiversity policy. Science and Public Policy 31 (2), 95–105. Environment Agency, 2009. Wild Salmon and Migratory Trout Tagging and Log Book Byelaws. . Environment Agency, 2005. Fisheries Statistics 2005. . Environment Agency and CEFAS, 2009. Annual Assessment of Salmon Stocks and Fisheries in England and Wales 2008. EA, Bristol. Finlayson, A.C., 1994. Fishing for Truth: A Sociological Analysis of Northern Cod Stock Assessments from 1977 to 1990. Institute of Social & Economic Research, Memorial University of Newfoundland, St. John’s. Franklin, A., 2001. Neo-Darwinian leisures, the body and nature: hunting and angling in modernity. Body & Society 7 (4), 57–76. Frear, P., 2008. Analysis of Angling Match Returns 2007 (Environment Agency North East Region) Dales Area Fisheries. EA, York.

S. Eden / Geoforum 43 (2012) 1014–1023 Greenwood, J.D., 2003. The monitoring of British breeding birds: a success story for conservation science? Science of the Total Environment 310, 221–230. Hannah, M.G., 2001. Sampling and the politics of representation in US Census 2000. Environment and Planning D: Society and Space 19, 515–534. Hinchliffe, S., 2008. Reconstituting nature conservation: towards a careful political ecology. Geoforum 39, 88–97. Kuklick, H., Kohler, R.E., 1996. Introduction. Osiris 11, 1–14. Latour, B., 1987. Science in Action. Open University Press, Milton Keynes. Latour, B., 1999. Pandora’s Hope. Harvard University Press, Cambridge, MA. Law, J., 2008. On sociology and STS. Sociological Review 56 (4), 623–649. Lawrence, A., 2006. No personal motive? Volunteers, biodiversity and the false dichotomies of participation. Ethics, Place and Environment 9, 279–298. Lee, M., 2009. Analysis of Angling Match Returns 2007 (Environment Agency North East Region) Yorkshire Area Fisheries. EA, York. Lorimer, J., 2006. What about the nematodes? Taxonomic partialities in the scope of UK biodiversity conservation. Social & Cultural Geography 7 (4), 539–558. Lulka, D., 2004. Stabilizing the herd: fixing the identity of nonhumans. Environment and Planning D: Society and Space 22, 439–463. Mansfield, B., 2003. From catfish to organic fish: making distinctions about nature as cultural economic practice. Geoforum 34, 329–342.

1023

Meyer, M., 2010. Caring for weak ties – the natural history museum as a place of encounter between amateur and professional science. Sociological Research Online 15 (2), 9. Murray, G., Neis, B., Palmer, C.T., Schneider, D.C., 2008. Mapping cod: fisheries science, fish harvesters’ ecological knowledge and cod migrations in the northern Gulf of St. Lawrence. Human Ecology 36, 581–598. Snyder, S., 2007. New streams of religion: fly fishing as a lived, religion of nature. Journal of the American Academy of Religion 75 (4), 896–922. St. Martin, K., 2005. Mapping economic diversity in the First World: the case of fisheries. Environment & Planning A 37, 959–979. Star, S.L., Griesemer, J.R., 1989. Institutional ecology, ‘translations’ and boundary objects: amateurs and professionals in Berkeley’s Museum of Vertebrate Zoology, 1907-39. Social Studies of Science 19, 387–420. UK National Ecosystem Assessment, 2011. The UK National Ecosystem Assessment: Synthesis of the Key Findings. UNEP-WCMC, Cambridge. Waterton, C., 2002. From field to fantasy: classifying nature, constructing Europe. Social Studies of Science 32 (2), 177–204. Waterton, C., 2003. Performing the classification of nature. In: Szerszynski, B., Heim, W., Waterton, C. (Eds.), Nature Performed. Blackwell, Oxford, pp. 111–129.