Adapting to the impacts of global change on an artisanal coral reef fishery

Ecological Economics 102 (2014) 118–125

Contents lists available at ScienceDirect

Ecological Economics journal homepage: www.elsevier.com/locate/ecolecon

Analysis

Adapting to the impacts of global change on an artisanal coral reef fishery Iliana Chollett a,b,⁎, Steven W.J. Canty c, Stephen J. Box d, Peter J. Mumby b,a a

Marine Spatial Ecology Lab, College of Life and Environmental Sciences, University of Exeter, Exeter EX44PS, UK Marine Spatial Ecology Lab, School of Biological Sciences, University of Queensland, Brisbane 4072, Australia c Centro de Estudios Marinos, Tegucigalpa, Honduras d Smithsonian Marine Station, Fort Pierce, FL 34949, USA b

a r t i c l e

i n f o

Article history: Received 13 November 2013 Received in revised form 10 March 2014 Accepted 18 March 2014 Available online 27 April 2014 Keywords: Small-scale fisheries Fuel price Wave exposure Climate change Emission scenarios Economic scenarios Adaptation

a b s t r a c t When assessing future changes in fishing, research has focused on changes in the availability of the resource. Fishers' behaviour, however, also defines fishing activity, and is susceptible not only to changes in weather but also to changes in the economy, which can be faster and more ubiquitous. Using a novel modelling approach and spatially explicit predictors we identified the current drivers of artisanal fishing activity and predicted how it is likely to change in 2025 and 2035 under two climate and two economic scenarios. The model is effective at explaining the activity of fishers (AUC = 0.84) and suggests that economic variables overwhelm the importance of climate variables in influencing the decisions of fishers in our case study area (Utila, Honduras). Although future changes in the overall incidence of fishing activity are modest, decreases in the number of accessible fishing grounds with projected increases in fuel prices will increase localised fishing effort depleting fish resources near the port. Compelling adaptation strategies in the area require the intervention of the market chain to make the sale price of fish more responsive to fuel price fluctuations and changes in fishing behaviour to improve fuel efficiency, including the revival of traditional ways of fishing. © 2014 Elsevier B.V. All rights reserved.

1. Introduction The artisanal fishing sector provides about 90% of all fishing employment worldwide and nearly 25% of world catches (Schorr, 2005). Artisanal fishing is, however, an economically risky activity, due to high livelihood and income dependency on local resources and the unpredictability of their availability. This makes the activity highly vulnerable to natural fluctuations in fisheries stock, its over-exploitation, and changes in economic drivers such as fuel prices and the supply and price of fish (Allison and Ellis, 2001). Changes in weather and the economy have the potential to modify fishing activity through changes in the behaviour of fishers and target abundance (Fig. 1). A number of key changes in the Earth's atmosphere and ocean have been detected in recent decades and associated with long-term environmental processes and human-induced factors (Solomon et al., 2007; Tett et al., 1999; Zhang et al., 2007). These include increasing global surface temperature, rising sea levels and changes in precipitation, as well as increases in the variability and intensity of extreme weather events (Emanuel, 2013; Held and Soden, 2006; Solomon et al., 2007). Although research on the effects of climatic change

⁎ Corresponding author at: Marine Spatial Ecology Lab, College of Life and Environmental Sciences, University of Exeter, Exeter EX44PS, UK. Tel.: +61 7 3365 1671. E-mail addresses: [email protected] (I. Chollett), [email protected] (S.W.J. Canty), [email protected] (S.J. Box), [email protected] (P.J. Mumby).

http://dx.doi.org/10.1016/j.ecolecon.2014.03.010 0921-8009/© 2014 Elsevier B.V. All rights reserved.

on fisheries is generally focused on the biological system (e.g. Brander, 2007; Cheung et al., 2010), environmental changes are also likely to affect the behaviour of fishers and their access to resources, a consideration that has been consistently neglected in climate change fisheries research (Haynie and Pfeiffer, 2012). Weather variables such as rainfall, strong winds or storms can restrict the number of fishing days (Lopes and Begossi, 2011), and exposure to waves determine accessibility, particularly for small vessels (Bastardie et al., 2013), with areas that are too rough being unsuitable for fishing (Hilborn and Kennedy, 1992). Climatic change signals are spatially heterogeneous, with some regions showing strong responses and others being almost untouched (Solomon et al., 2007). Future changes in the economy, however, are more ubiquitous (EIA, 2013), and economic constraints have the potential to overwhelm the decisions made by fishers. In selecting a location in which to fish, rational fishers consider not only the catch that they are likely to make, as judged by contemporary information and their own experience, but also the difficulty and costs of fishing under the conditions of a particular day at a particular location. The relevance of economic constraints in the decision-making process of fishers has been considered (e.g. Bastardie et al., 2013; Hilborn and Kennedy, 1992), but projections of economic drivers (e.g. EIA, 2013) have rarely been used when assessing possible changes in fishing activity over time. Where fisheries research has focused on fishing activity as a function of weather, analyses of the influence of economic drivers had tended to occur separately (Haynie and Pfeiffer, 2012; Johnson et al., 2013; but see Wilen et al., 2002). Here we assess the effects of both drivers on

I. Chollett et al. / Ecological Economics 102 (2014) 118–125

119

Economy Cost of fishing

Expected catch value

Weather

Storms Wave exposure

Fishing activity

Catch

Target biology

Target abundance

Rainfall Water temperature

Fig. 1. Conceptual model of linkages between weather and economic drivers of fishing activity, target biology and ultimately fishing catches.

artisanal fishing activity simultaneously, and predict future changes under different climate and economic scenarios. The novelty of this research includes: (1) the spatially-explicit quantification of wave exposure, which restricts fishing access; (2) the use of a novel modelling approach (MaxtEnt, a presence-only model) because of the occurrence of unreliable zeros in our dataset, which is a common problem in fisheries data; (3) the use of downscaled climate change models to forecast region-specific weather changes; and (4) the use of economic models to forecast changes in fishing costs. We apply the model to an artisanal fishery on the island of Utila in the Bay Islands (Honduras), where weather seems to be particularly relevant in shaping fishing activity. By understanding current drivers of fishing activity and how these are likely to change in the future, we hope to gain insight into how fisheries are likely to change and provide advice on building adaptive capacity in this vulnerable system. 2. Material and Methods 2.1. Study Area This study was carried out in Utila, Bay Islands, in the Honduran Caribbean (Fig. 2). The Bay Islands have a tropical climate with a wet season that runs from October to January, an average monthly rainfall maximum of about 10.44 mm day− 1, and daily maximum rainfall reaching up to 22.8 mm day− 1. Winds come from the east with a mean direction of 75.5 ± 51.7°, at an average speed of 4.6 ± 2.4 ms−1 (average and standard deviation, inset in Fig. 2). There are no large seasonal shifts in wind conditions though winds are weaker between September and January. Wave exposure closely follows the wind patterns. The eastern (windward) side of the islands is heavily exposed, while the west (leeward) side is generally protected from the action of wind-induced waves (Fig. 2). Artisanal fishers in Utila can access four distinct fisheries including hand-lining for coral reef associated fin fish or deep shelf snapper species, trolling for pelagic fish species or the collection of lobster and conch using SCUBA equipment. The fishing community that populates two small cays southwest of Utila is the most active in the Bay Islands, most heavily reliant on fishing (60% of the households are directly reliant on fishing: Box and Canty, 2010) and fish over the largest area

(Gobert et al., 2005). Fishers use small boats (up to 38 ft) with inboard diesel engines of about 15–70 horse power (pers. obs). Hand lines are the dominant fishing gear, but fishers might also use traps to target grouper (Serranidae) spawning aggregations and occasionally mesh nets when king mackerel (Scomberomorus cavalla) migrate close to the island. Shallow, coral reef associated fish species are the most important targets for the hand-line fishery, with the main species being yellowtail snapper (Ocyurus chrysurus), accounting for 36% (by weight) of the total landed catch (Box and Canty, 2010). Fisheries in Utila work on a share system: the fishers are not paid a wage but instead they receive a share of the profit after the costs of fuel, bait and ice have been deducted from the total catch revenue. Boats normally have two fishers who work together for a common catch. The skipper receives 35% of the profit while the second man receives 30%. The owner of the fishing boat receives 35% of the profit and is responsible for the maintenance of the vessel. Owners may skipper their own vessel or lease their boat to a different skipper under this payment system. If there is no catch fishing, it is the responsibility of the skipper to cover the fishing costs. Additional fishers may join a fishing pair under a separate fishing agreement. Additional fishers contribute around 1/5 of the total fuel costs and provide their own ice, bait and tackle. They keep their catch separate to the main fishing pair and on returning to port give half of their total catch to the skipper as payment. Fuel comprises about 75% of the operating costs of the trip, with average fuel consumption in the order of 4 gal of diesel at about USD 14.72 in 2010. Although the fishery involves few associated operating costs, they take about 50% of the gross revenue of the boat in an average trip. As a result, the fishery relies in high yield events of spawning aggregations or migrations to elevate the income (Box and Canty, 2010). The Honduran government's capacity for monitoring, control and surveillance of the artisanal fishery is weak. Artisanal fishing vessels, classified as those under three metric tonnes, are required to be centrally registered and licenced annually, as are artisanal fishers. These prerequisites for fishing are poorly enforced. There are no regulations controlling the minimum size for fin fish caught by artisanal fishers and no limits on catch or effort. Species-specific protection is afforded only to Nassau grouper (Epinephelus striatus) at their seasonal spawning aggregations and the retention and sale of all shark species are prohibited. Regulations exist which prohibit the use of poisons,

120

I. Chollett et al. / Ecological Economics 102 (2014) 118–125

16.4° N 330

0

Fishing grounds

90

270 240

16.2° N

Roatan

60

300

Legend Port

30

210

120 180

Roatan airport

150

Wave exposure (Jm-3 ) 1000 500 0

16.0° N Relative location

Cayos Cochinos

15.8° N La Ceiba

87.6° W

87.4° W

87.2° W

-87.0° W

86.8° W

86.6° W

0

12.5

25 km

86.4° W

Fig. 2. Fishing grounds around Utila (Bay Islands) and average wave exposure for the study area. The wind rose in the upper left corner indicates the overall wind regime in the area, dominated by easterly trade winds (in situ data for Roatán airport, 2000–2012).

explosives, beach seines and stipulate a minimum mesh size for gill nets, with further gear restrictions inside designated marine protected areas. Enforcement however is inconsistent across the Honduran Caribbean. There are currently no policies for fishery development (e.g. loans, subsidies) and catch data are not recorded by the government for the artisanal fisheries of the Bay Islands.

2.2. Data Sources Daily records of fishing activities for the period August 2010– September 2011 were recorded at the Utila Cays (16.0639°N, 86.9667°W). The 1,186 landing records collected include data on the date, fisher name, departure time and coordinates of the fishing ground visited. Data collection was organised by the Honduran Nongovernmental Organisation Centro de Estudios Marinos (CEM) and carried out at the home port by a local female. Fishers in Utila use a total of 39 fringing reefs and banks up to 60 km away from the port (Fig. 2). Daily in situ rainfall and wind speed data for Roatán airport (16.31°N, 86.53°W, Fig. 2) were obtained for the period 2000–2012 from wunderground.com and tutiempo.net. Data for the target period (August 2010–September 2011) were typical when compared to the climatology. Climate change projections were based on the ECHAM4 general circulation model downscaled to 40 km, considering two emission scenarios (SRES A2 and B2) for 2025 (mean of 2024–2026) and 2035 (mean of 2034–2036). Predicted rain and wind data were obtained from the Caribbean Community Climate Change Centre. The Centre produces downscaled climate projections using the PRECIS (Providing Regional Climates for Impacts Studies) software developed at the Met Office Hadley Centre (Jones et al., 2004). The scenarios represent different paths of future economic development and energy use (Nakicenovic and Swart, 2000). The SRES A2 is a ‘business as usual’ emission scenario, projecting 3 °C increase in surface air temperature by 2100. The scenario SRES B2 is less energy-intensive and features a low emission path, projecting a 2.2 °C temperature increase, and it has been related to a future ‘green economy’. PRECIS outputs are monthly percentage of change in weather conditions. Projected future daily data for the years 2025 and 2035 were estimated by adding the observed August 2010–September 2011 daily data and the amount of change for the corresponding month and year. Although the temporal resolution of the projections is monthly, the outputs were kept at daily resolution to retain the extremes in the data.

Utila fishers use diesel as fuel. Historical data on diesel prices for Honduras were obtained from the GIZ International Fuel Prices Database. Annual prices of diesel are correlated to prices of crude oil (Brent barrel) described by the U.S. Energy Information Administration (EIA, Diesel = 0.1912 + 0.0081 ∗ Oil, R2 = 0.90, n = 9). We used this relationship, alongside forecasted crude oil prices from the EIA to assess future changes in diesel prices (in USD) under two different economic scenarios: a Reference Oil Price case (ROP) and a High Oil Price case (HOP, EIA, 2013). The scenarios were developed by adjusting the economics of the supply, investment, and production of oil in the future as well as economic growth of the main countries involved. 2.3. Wave Exposure Calculation Wave exposure maps describe the sea condition and the degree of wave action on an open shore, and are therefore a proxy for site accessibility, with rough, high wave exposure areas being inaccessible to most fishers. Daily wave exposure maps were calculated following the methods described in Chollett and Mumby (2012) using a wave theory method where exposure is a function of the shape of the basin (i.e. fetch), wind speed and direction. Coastline data were produced from Landsat satellite imagery. Fetch in 64 compass directions was calculated by tracing a line from each marine location (pixel) across the sea until land was encountered. We used in situ wind speed and direction data (see previous section) for the target period (August 2010–September 2011) to calculate daily exposure at a spatial resolution of 50 m. Detailed equations can be found in Ekebom et al. (2003) and Chollett and Mumby (2012). 2.4. Modelling Fishing Activity The response variable used was the presence of fishers at each fishing ground on each particular day, providing insight into fishers' short-term decisions. To explain fishing activity, we considered a predictor set comprised by eight variables summarising information on weather, cultural, and economic aspects of the fisheries. Weather was included in the form of current (on the day) and changed conditions (difference between the value of the day and the average of previous three days) for wind speed (m s−1) and rainfall (mm). The rationale of including variables for changed conditions was that if unfavourable fishing weather conditions were present in the immediate past, it might be more likely for fishers to go fishing. Wave exposure at the fishing ground (Joules m−3) was included as a site-specific variable, as well as the distance to the fishing ground, which has economic implications since greater distances require more fuel. We calculated path distances, as opposed to linear distances,

I. Chollett et al. / Ecological Economics 102 (2014) 118–125

between the home port and the fishing ground to account for the presence of islands that interfere in the route of the boat, thereby increasing distances travelled. Day of the week (categorical: weekday or weekend) was a cultural variable added to account for the possible lack of fishing during religious breaks. Finally, fishing around Utila occurs during morning and evening hours, so this covariate was also included (categorical: morning or afternoon). 2.5. Modelling Approach Our dataset presented several issues that made it unusable for common modelling approaches such as generalised linear or additive models. The dataset includes a large number of zeros and there were many instances where appropriate conditions for fishing were present but no fishing was recorded. We considered these zeros unreliable based on three facts: (1) the number of fishing grounds in Utila is large relative to the number of fishers so there is a lack of saturation; (2) fishers often cluster at specific sites, leaving many other sites unvisited; and (3) most importantly, we detected observer inconsistency, an unfortunate but common issue when data needs to be collected on a continuous basis by a human observer. We solved this modelling challenge using a Maximum Entropy Model (MaxEnt), a presence-only modelling method that has been widely used in ecological and conservation applications to predict current and expanding distributions of species. Presence-only models are a type of modelling approach that is recommended when absence data are not available, are meaningless, or unreliable (Reutter et al., 2003). MaxEnt has shown superior predictive performance over comparable methods (Elith et al., 2006; Elith et al., 2011; Wisz et al., 2008). A detailed description of the machine learning method can be found in Phillips and Dudík (2008), Phillips et al. (2006) and Elith et al. (2011). In brief, MaxEnt uses information on the conditions in the region of interest (the background) as a basis for comparison with conditions at known presence sites. The software computes a probability density distribution for the background, and then attempts to find a probability distribution for ‘presences’ which fulfils two requirements: (1) the expected value of the explanatory variable is within a confidence interval of its mean over the presence records; and (2) within all the distributions that satisfy the constraints, the programme chooses the probability distribution of maximum entropy (most spread out: Phillips et al., 2006). MaxEnt allows for non-linear functions to be fitted to each predictor and pairwise interactions. The background dataset included areas that are potentially available for fishing (i.e. in which fishing could have occurred if conditions were suitable). Here, we used as background all data for fishing grounds where fishing was and was not reported (total of 34,348 records). The predictor set included all records where fishing occurred (1,186 records). Our location-based (non-gridded) data was modelled using the samples-with-data format in MaxEnt. Default settings for features and regularisation were used for model training, and 10-fold cross validation was used to obtain estimates of predictive performance and uncertainty around fitted functions. All analyses were carried out using MaxEnt version 3.3.3 (At&T Labs, Florham Park, New Jersey, USA). We explored the initial dataset in order to remove highly correlated (and therefore redundant) variables before the analyses. Model refinement was done using the jackknife feature in MaxEnt to assess performance of each variable in terms of AUC (Area Under the Curve of Receiver–Operator Characteristic) gain. The AUC score provides a single measure of overall accuracy of the model that is independent of the threshold chosen to identify activity, making a better use of all the information provided by the classifier, and reflects the probability that a randomly chosen presence site will rank above a randomly chosen background site. AUC values range from 0.5 to 1. While an AUC score of 0.5 indicates randomness, one of 1.0 indicates perfect model performance (albeit never achieved in MaxEnt). Variables that decreased the predictive performance of the model were omitted from the final analysis.

121

Once the model was developed using historical information, we projected the model into the future by applying it to another set of predictors for two time periods: 2025 and 2035, which are relevant time frames for economic projections and fisheries management actions. We assessed changes in fishing activity under a changing climate and economy by using forecasted data on weather and fuel prices. Future daily wave exposure and rainfall were calculated according to climate change predictions as indicated Section 2.2. In this assessment, changes in economy and increases in fuel prices effectively decrease the number of fishing grounds available for fishing. For a skipper, decisions about where to fish are a trade-off between anticipated catch revenue and the cost of fuel to reach the site. Consequently, if the expected revenue (dockside prices) remains constant (which has been the case in Utila since 1992: Box and Canty, 2010), increases in the cost of the trip will decrease fishing activity at sites further from port until reaching a threshold distance of no profitability and no activity. The total cost of fuel per trip is related to the cost per litre of fuel and the amount of fuel required, which is determined by the distance travelled and the efficiency of the engine (about 0.85 l km−1). By using the observed distance threshold where fishing activity exhibits a marked decrease, and the known boat efficiency and fuel costs in 2010, it is possible to know the threshold cost of the trip above which fishers' activity drops. We then applied this cost threshold to projected fuel prices in 2025 and 2035 to estimate the maximum distances likely to be travelled under future economic scenarios. Using this information, fishing grounds beyond this threshold distance were excluded from the scenarios of future fishing conditions. In 2010 only 0.34% of the trips occurred above this threshold distance, showing that albeit simple, this method captures most of the complexity in the economic decisions taken by fishermen in Utila. 3. Results 3.1. Current Drivers of Fishing Activity The initial eight predictor variables were simplified to a final suite of five variables (Table 1). After data exploration, we excluded wind speed, which was correlated with wave exposure (r = 0.64); and rainfall change, which was highly correlated with the rainfall on the day (r = 0.81). After fitting the full model in MaxEnt excluding these two variables, we observed that the inclusion of wind speed change decreased the performance of the model (Table 1), and therefore this variable was also removed from the analysis. The final model performed well in terms of a high AUC score (84%). The overall probability of fishing activity was low (0.23 ± 0.20, average and standard deviation over all days and fishing grounds). This implies that on average, on any given day each particular fishing ground as a 2 in 10 chance of being visited. Distance to port was the most influential variable in the model, followed by the time of fishing. Weather and cultural variables had less importance in modulating fishing activity (Table 1). The marginal response curves for the variables included in the final model (Fig. 3) represent a simplification of the response, where interactions are not taken into account. However, some patterns are obvious such as fishing activity was associated with intermediate distances from port (5–40 km, Fig. 3a): any given fishing ground at less than 5 km from port has less than a 5% probability of being visited on any given day. Fishing grounds that are located farther offshore, at a distance of ~10–15 km, have a much higher chance of being visited (N70%). The probability of fishing activity then decreases for fishing grounds located even farther, particularly once distance exceeds 40 km, where the likelihood of being visited drops rapidly. Fishing activity is also more likely associated with lower rainfall (b10 mm day−1, Fig. 3c) and lower wave exposure (b2500 J m3, Fig. 3d). Although fishing is unlikely when extreme weather events occur, those episodes were rare in the region (histograms in Fig. 3c and d), which might explain the limited importance of weather variables in

122

I. Chollett et al. / Ecological Economics 102 (2014) 118–125

Table 1 Model AUC and percentage contribution of predictor variables to each model of fishing activity. Variable contribution (%) Distance

Fishing time

Wave exposure

Rainfall

Day of the week

Wind speed change

0.83 ± 0.01 0.84 ± 0.01

51.26 53.10

35.64 34.41

5.11 5.64

3.39 3.52

3.41 3.33

1.19

the model. This suggests that even if the relative contribution of these variables to the model is low (Table 1), they have the potential of limiting fishing activity if the weather conditions become increasingly unfavourable in the future. 3.2. Changes in Fishing Activity

a

0.7

b

0.5

50

15

0.5

0.3 30

0.3 5

0.1 0

10

20

30

40

50

0.1

10

am

60

Distance (km) p of fishing activity

70

frequency (%)

25

0.9

pm

Fishing time

0.7 90

c

0.5

d

0.5

70

70 50

0.3

50

0.3

30

30 0.1

10 20

40

60

80

0.1

10 2000

100

p of fishing activity

Rainfall (mm)

6000

10000

Wave exposure (J

e

70

50 30

30

10

10

14000

m-3)

70

50

weekday

90

frequency (%)

p of fishing activity

When looking at future changes in weather, results from the business as usual (A2) and green economy (B2) scenarios were very similar for the variables and time frames assessed, providing differences in relative changes to present day conditions of only 2.1% (2025) and 6.7% (2035) for wave exposure and 2.0% (2025) and 9.3% (2035) for rainfall. Therefore, for the sake of simplicity, we only comment on the results of the A2 scenario. On average, rainfall was projected to decrease by about 19.2% by 2025 and 39.2% by 2035. Whereas wave exposure is likely to increase by 7.5% in 2025 and decrease by 6.0% in 2035, following changes in wind intensity. Both economic scenarios predicted increases in fuel costs with time, therefore decreasing the number of available fishing grounds. Under the

High Oil Price (HOP) scenario, larger increases in fuel cost produced a larger reduction in the number of fishing grounds accessible to fishing. Reductions are in the order of 25.6% and 38.5% for 2025 and 2035 respectively. Under the Reference Oil Price (ROP) scenario the number of accessible sites decreased by 15.4% and 23.1% for both years. These 8 future global change scenarios (for every possible combination of the two climate and the two economic scenarios for the two time periods considered) and the model of fishing activity were used together to assess likely changes in future fishing activity using MaxEnt. All scenarios provided significantly different results to current conditions (ANOVA, F = 26.21, p b 0.001) and all showed increases in fishing activity, although the differences to present conditions are small (of just 1–2% on average, Fig. 4). When comparing only the future scenarios, the fishing activity predicted by each of them is also significantly different (ANOVA, F = 6.33, p b 0.001, Fig. 4). Not surprisingly, outputs with the business as usual and green economy climate scenarios offered similar results (Fig. 4, multiple comparison test, p N 0.05). The only scenarios that were significantly different were the 2025 Reference Oil Price and the 2035 High

frequency (%)

Full model Final model

AUC (average, s.d.)

weekend

Fishing day Fig. 3. Marginal response curves (hinges) showing how each explanatory variable affects the MaxEnt prediction and the probability of fishing activity. Black lines indicate the mean of 10 cross-validation replicates, and light grey lines (or error bars) the standard deviation across these replicates. The figure also shows in light grey the histograms for the background dataset (when fishing did and did not occur). (a) Distance from port; (b) fishing time; (c) rainfall on the day; (d) wave exposure at the site; (e) day of the week.

I. Chollett et al. / Ecological Economics 102 (2014) 118–125

Current

123

39

2025 A2 ROP

33

Scenario

2025 B2 ROP 2025 A2 HOP

29

2025 B2 HOP 2035 A2 ROP

30 2035 B2 ROP 2035 A2 HOP

24

2035 B2 HOP

0.229 0.230 0.232 0.234 0.236 0.238 0.240 0.242 0.244 0.246 0.248 0.250 0.252 0.254

Probability of fishing activity Fig. 4. Probability of fishing activity for current conditions and future scenarios for an average fishing ground. Mean and 95% confidence intervals are shown, as well as the number of available fishing grounds for each scenario.

Oil Price scenarios (Fig. 4, multiple comparison test, p b 0.05) which represent extreme cases of future conditions. Differences were however, very small, less than 1% in average probability of fishing activity. MaxEnt predictions indicate that fishing activity is likely to increase slightly in the future. These increases in activity are mainly related to increases in fuel prices and not related to changes in climate. Climate change scenarios forecast more favourable conditions for fishing in 2035 compared to 2025 (lower exposure and rain). Results from the 2025 High Oil Price and 2035 Reference Oil Price scenarios, where fuel price constraints in the number of fishing grounds available are comparable (29 and 30 respectively), are however indistinguishable (Fig. 4), indicating that weather has a limited influence in defining future fishing activity in the region. Increases in the cost of fuel will encourage the use of fishing grounds closer to the home port, which currently have a high associated probability of activity (Fig. 3a), therefore increasing the overall probability of activity in the future. Although this might seem positive at a first glance, it would lead to the spatial concentration of fishing effort at a reduced number of fishing grounds and increased competition between fishers at these sites, which more likely will deplete the resource at those locations. 4. Discussion In areas where the weather is mild, fishing activity is defined by economic rather than weather constraints. Economic aspects are common controls of small-scale fisheries with a high dependency on catch rates (De Camargo and Petrere, 2001) such as the Bay Island fisheries (Box and Canty, 2010). The relevance of economic constraints for this particular case study is explained by the tight dependency on the cost of fuel, the only direct cost assumed by the skipper, who makes the decisions on when and where to go fishing, and the fact that this fishery has limited profitability and high reliance in high yield events such as spawning aggregations or migrations (Box and Canty, 2010). Wave exposure, a variable rarely assessed in a fisheries context (but see Hilborn and Kennedy, 1992) and not previously assessed in a spatially-explicit manner, was found to be the most relevant weather variable explaining the decisions of fishers. The lack of importance of other weather predictors in the Bay Islands was related to the relatively mild and stable weather of the area. In Utila, weather conditions unsuitable for fishing, such as high exposure (N2500 J m−3) and heavy rainfall (N10 mm) are rare, occurring 2 and 10% of the time respectively. We could expect that the relative importance of weather predictors will be higher in regions where weather conditions are harsher (Hilborn

and Kennedy, 1992) particularly in fisheries with small vessels that are more vulnerable to the weather factor (Bastardie et al., 2013). Downscaled weather predictions for the Bay Islands showed improved weather for fishing in the future, effectively opening opportunities for fishing. These weather projections coincide with results of other modelling approaches for the region (e.g. Campbell et al., 2011). Changes in weather are, however, spatially heterogeneous, and some areas in the Caribbean are forecasted to become wetter (Campbell et al., 2011), which could limit small-scale fishing activity in the future. Sitespecific assessments therefore need to be done in order to assess the vulnerability of different fisheries to climate change and the necessity of implementing climate change adaptation measures into fisheries management plans. The spatial distribution of fishing effort is also likely to change because of changes in the spatial distribution of the stock. Climate change may influence stock abundance through changes in individual performance (i.e. growth, survival and reproduction), biological interactions, population connectivity or habitat loss (Munday et al., 2008). Climate change has been associated with shifts in the distribution of fisheries resources, mainly related to increases in temperature and associated latitudinal (pole-ward) or bathymetric (to deeper areas) shifts (e.g. Perry et al., 2005) or with dramatic changes in species abundances (Simpson et al., 2011) and the production of novel fish assemblages (Harborne and Mumby, 2011). The nature of these changes in distribution would be dependent on the rates of climate change in the area, the temperature tolerance of the species involved and the availability of suitable habitat. Most species of commercial importance in the Bay Islands are broadly distributed and their geographical distributions might not be affected by climate change in the time frames considered here. The interaction of climate change effects in the stock and the behaviour of the fish harvesters, however, should be considered when assessing longer timeframes. The artisanal fisheries of the Bay Islands were characterised in 2005 as a targeted fishery with moderate fishing pressure and limited impacts on habitat (Gobert et al., 2005). The depletion of large bodied, highly valued snapper and grouper species (Gobert et al., 2005); the transition in landings to lower trophic level species, including smaller and faster growing yellowtail snapper as well as the diversification of the fishing techniques to include the use of traps and nets are all signs of overexploitation of the coral reef fishery. The current study has also shown lower fishing effort on fishing sites close to the port, indicating spatial variability in the overexploitation of the remaining species. Although many factors come into play in the decision of the fishermen

124

I. Chollett et al. / Ecological Economics 102 (2014) 118–125

on how far to travel, such as their familiarity with the fishing grounds, competition with nearby fishing communities, or safety issues (Daw, 2008), the spatial behaviour of the fishermen in Honduras can be accurately explained by a simple model only based on fuel costs and resource availability (Caddy and Carocci, 1999). The pattern of changes in fishing activity with distance to the port follows what Caddy and Carocci (1999) called the ‘Gaussian effort allocation model’: during the early stages of exploitation by a local artisanal fishery the fishing grounds closer to the port are more attractive because they incur less travelling time and fuel expenses. These resources are soon depleted, attracting lower and lower effort despite their accessibility. With time, the peak of effort with distance is progressively transferred further away as the stock is harvested and the fishery progressively occupies a broader envelope of space with time (Caddy and Carocci, 1999). This potential for expansion is, however, limited by fuel costs, as an openaccess fishery will continue to expand its fishing grounds until the increased fuel costs offset the revenue provided by the new fish stocks. Expansion may also be limited by territorial limits, such as the existence of marine protected areas, traditional divisions of fishing grounds between neighbouring communities or international maritime boarders. Some have suggested that increasing fuel costs will help reduce overfishing, because this reduces the profitability of the fishing activity causing fishers to leave the fishery (Arnason, 2007; Sumaila et al., 2008), which would in turn reduce excessive fishing pressure leading to some fish stock recovery and the possibility of higher catches with less effort in the future (Arnason, 2007). Increasing fishing costs however are also an important driver for illegal fishing activity including the incursion of fishers into marine protected areas or poaching from the territorial waters of neighbouring countries (Perez, 2009). In the region and time frames considered here, it seems more likely that increases in fuel prices will lead to increasing spatial competition between fishers and increasing rates of resource depletion near the port as has been reported in other areas under similar conditions (Abernethy et al., 2010). There is also a greater likelihood of fishers adopting less selective fishing methods to increase catch volume, or to break spatial management regulations such as fishing inside no-take reserves. Although increases in fuel prices might decrease fishing effort in the long term, they will likely have social and ecological consequences in the short to mid-term. The establishment of a data collection system to monitor fish resources is desirable, as well as the development of participatory, community based management where the fishers are directly involved in the collection and use of fisheries data to inform management strategies that can ensure the sustainability of the fishery they depend on. Although fuel subsidies are not currently provided in Honduras, they are a common means of alleviating social concerns over increased fuel prices (Sumaila et al., 2008). However, subsidies are unsustainable in the long-term, because they are expensive, discourage the retirement of fishers with negative or marginal profits, lower the retail prices of the catch, increase fishing effort and decrease sustainability (Markus, 2010; Sumaila et al., 2008). To offset the increased fuel costs, another alternative is the achievement of fuel-efficient fishing either through behavioural changes, technological improvements or the revival of traditional ways of fishing. Unlike fuel subsidies, these alternatives have the benefit of encouraging the retirement of unprofitable fishers, and most importantly, of avoiding the artificial decrease of the price of the catch which will in turn increase the demand of a product that is already under pressure. This strategy, however, has the potential of increasing fishing effort and decrease sustainability, just as subsidies do. Therefore, if adopted, it needs to be part of a management strategy that not only promotes fuel efficiency, but low impact fishing (Suuronen et al., 2012) to promote the sustainable use of fisheries resources. Low impact, fuel-efficient fishing should encourage the use of passive fishing gears (e.g. hooks instead of nets) which reduce ecosystem impacts and carbon footprint of the activity. Changes in fishing behaviour include reducing vessel speed when travelling to grounds, increasing the length of the trip to increase fishing

time relative to travel time or limiting fishing to optimal weather conditions (Abernethy et al., 2010; Suuronen et al., 2012). Fuel efficiency is related to weather, with rough conditions increasing the fuel consumption of vessels, which waste energy breaking the waves (Bastardie et al., 2013). Projected decreases in wind speed and wave exposure in the study area might be associated to more fuel-efficient conditions, a factor that was not considered in this study. Technological improvements to achieve fuel efficiency include the modernisation and renewal of the fishing fleet in the short term, and experimentation with alternative energy technologies such as compressed air engines or biofuels (Suuronen et al., 2012) in the mid to long term. Another possibility would be to undertake the revival of traditional ways of fishing that use paddle or wind assisted propulsion (or a combination of the two), which remains a common alternative for fishers in other parts of the Mesoamerican Barrier Reef system (Huitric, 2005), and has been shown to be a more profitable choice in small artisanal fisheries with fluctuating catches (De Camargo and Petrere, 2001). While decreasing operating costs will help maintain a profitable fishery, an alternate adaptation strategy under a changing economy is to increase the gross revenue of the fishing trip by improving the price of fish at the first point of sale. Fish market prices have remained stagnant for the past 10 years in the Bay Islands in spite of increases in fish retail prices (Box and Canty, 2010), a phenomenon that has also been reported in Europe (Abernethy et al., 2010). Stationary fish prices are related to local or global market constraints which determine the price of the catch (Abernethy et al., 2010). In the Bay Islands the obstacle is local: fish are landed locally and bought by two buyers which maintain a constant price. The lack of storage capacity for fresh fish by individual fishers and the high cost of transporting the fish away from the island force the fishers to sell locally and absorb the increasing fishing costs (Box and Canty, 2010). The institutional set-up of the retail chain should be reviewed in the Bay Islands so that prices are more responsive to fishing costs and fishers receive a better price for their catch, which would in turn alleviate the pressure on local resources. Increasing the gross revenue of the activity, however, can be a double-edge sword because long-term increase in prices in open-access regimes may trigger exponential increases in fishing effort and the collapse of the fishery, as has occurred in several artisanal fisheries in Latin America with the globalisation of markets (Defeo and Castilla, 2012; Defeo et al., 2013). To avoid this prospect, the governance system in the Bay Islands should be strengthened through the development of genuine community based co-management (Basurto et al., 2012; Gutierrez et al., 2011). Successful co-management arrangements have been related to both social and ecological benefits for the fishery, such as community empowerment, increased collaboration and learning between partners, integration of scientific and local knowledge, high levels of compliance and increases in fisheries stocks (Cinner et al., 2012; Kittinger et al., 2013). Although co-management agreements are in place in this region of Honduras, these merely function as mechanisms for the delegation of management responsibilities to Non-governmental Organizations or Civil Society Organizations which do not represent the interests of local stakeholders. It is a priority to facilitate the development of true co-management arrangements with effective stakeholder participation, strong leadership and social cohesion to guarantee the sustainability of the fisheries in the region (Gutierrez et al., 2011). Understanding historical drivers of fishing activity is essential to predict the future behaviour of fisheries and plan for changes in climate and the economy. In Honduras, location-specific fishing costs and weather conditions play important roles in driving fisher behaviour. While changes in climate are unlikely to have major impacts on the fishery, increases in fuel prices will likely have a deleterious impact on profit, and therefore require a shift in fishery governance and technology in order to maintain livelihoods. This study provides a framework for the assessment of global changes in fishing activity, in areas with contrasting weather regimes and economic dependencies.

I. Chollett et al. / Ecological Economics 102 (2014) 118–125

Acknowledgements The research leading to these results has received funding from the European Union 7th Framework programme (P7/2007-2013) under grant agreement no. 244161, Pew (grant 2008-000330-010) and ARC Laureate Fellowships (FL0992179) to PJM and the Summit Foundation (grant 504502) and Smithsonian Marine Station at Fort Pierce contribution number 949 to SJB. We are grateful to the Caribbean Community Climate Change Centre, particularly Timo Baur, for assisting in the provision of climate change data, and to Alice Rogers and two reviewers for their comments on the manuscript. References Abernethy, K.E., Trebilcock, P., Kebede, B., Allison, E.H., Dulvy, N.K., 2010. Fuelling the decline in UK fishing communities? ICES J. Mar. Sci. 67, 1076–1085. Allison, E.H., Ellis, F., 2001. The livelihoods approach and management of small-scale fisheries. Mar. Policy 25, 377–388. Arnason, R., 2007. The economics of rising fuel costs and European fisheries. EuroChoices 6, 22–29. Bastardie, F., Nielsen, J.R., Andersen, B.S., Eigaard, O.R., 2013. Integrating individual trip planning in energy efficiency — building decision tree models for Danish fisheries. Fish. Res. 143, 119–130. Basurto, X., Cinti, A., Bourillón, L., Rojo, M., Torre, J., Weaver, A.H., 2012. The emergence of access controls in small-scale fishing commons: a comparative analysis of individual licenses and common property-rights in two Mexican communities. Hum Ecol 40, 597–609. Box, S.J., Canty, S.W.J., 2010. The long and short term economic drivers of overexploitation in Honduran coral reef fisheries due to their dependence on export markets. 63rd Gulf and Caribbean Fisheries Institute Conference, San Juan, Puerto Rico, pp. 43–51. Brander, K.M., 2007. Global fish production and climate change. Proc. Natl. Acad. Sci. 104, 19709–19714. Caddy, J.F., Carocci, F., 1999. The spatial allocation of fishing intensity by port-based inshore fleets: a GIS application. ICES J. Mar. Sci. 56, 388–403. Campbell, J.D., Taylor, M.A., Stephenson, T.S., Watson, R.A., Whyte, F.S., 2011. Future climate of the Caribbean from a regional climate model. Int. J. Climatol. 31, 1866–1878. Cheung, W.W.L., Lam, V.W.Y., Sarmiento, J.L., Kearney, K., Watson, R.E.G., Zeller, D., Pauly, D., 2010. Large-scale redistribution of maximum fisheries catch potential in the global ocean under climate change. Glob. Chang. Biol. 16, 24–35. Chollett, I., Mumby, P., 2012. Predicting the distribution of Montastraea reefs using wave exposure. Coral Reefs 31, 493–503. Cinner, J.E., McClanahan, T.R., MacNeil, M.A., Graham, N.A.J., Daw, T.M., Mukminin, A., Feary, D.A., Rabearisoa, A.L., Wamukota, A., Jiddawi, N., Campbell, S.J., Baird, A.H., Januchowski-Hartley, F.A., Hamed, S., Lahari, R., Morove, T., Kuange, J., 2012. Comanagement of coral reef social-ecological systems. Proceedings of the National Academy of Sciences 109, 5219–5222. Daw, T.M., 2008. Spatial distribution of effort by artisanal fishers: exploring economic factors affecting the lobster fisheries of the Corn Islands, Nicaragua. Fish. Res. 90, 17–25. De Camargo, S.A.F., Petrere Jr., M., 2001. Social and financial aspects of the artisanal fisheries of Middle São Francisco River, Minas Gerais, Brazil. Fish. Manag. Ecol. 8, 163–171. Defeo, O., Castilla, J.C., 2012. Governance and governability of coastal shellfisheries in Latin America and the Caribbean: multi-scale emerging models and effects of globalization and climate change. Current Opinion in Environmental Sustainability 4, 344–350. Defeo, O., Castrejón, M., Ortega, L., Kuhn, A.M., Gutiérrez, N.L., Castilla, J.C., 2013. Impacts of climate variability on Latin American small-scale fisheries. Ecology and Society 18. EIA, 2013. International Energy Outlook 2013. OECD Publishing, Washington. Ekebom, J., Laihonen, P., Suominen, T., 2003. A GIS-based step-wise procedure for assessing physical exposure in fragmented archipelagos. Estuar. Coast. Shelf Sci. 57, 887–898. Elith, J., Graham, C.H., Anderson, R.P., Dudik, M., Ferrier, S., Guisan, A., Hijmans, R.J., Huettmann, F., Leathwick, J.R., Lehmann, A., Li, J., Lohmann, L.G., Loiselle, B.A., Manion, G., Moritz, C., Nakamura, M., Nakazawa, Y., Overton, J.M., Peterson, A.T., Phillips, S.J., Richardson, K., Scachetti-Pereira, R., Schapire, R.E., Soberon, J., Williams, S., Wisz, M.S., Zimmermann, N.E., 2006. Novel methods improve prediction of species' distributions from occurrence data. Ecography 29, 129–151. Elith, J., Phillips, S.J., Hastie, T., Dudík, M., Chee, Y.E., Yates, C.J., 2011. A statistical explanation of MaxEnt for ecologists. Divers. Distrib. 17, 43–57.

125

Emanuel, K.A., 2013. Downscaling CMIP5 climate models shows increased tropical cyclone activity over the 21st century. Proc. Natl. Acad. Sci. 110, 12219–12224. Gobert, B., Berthou, P., Lopez, E., Lespagnol, P., Turcios, M.D.O., Macabiau, C., Portillo, P., 2005. Early stages of snapper–grouper exploitation in the Caribbean (Bay Islands, Honduras). Fish. Res. 73, 159–169. Gutierrez, N.L., Hilborn, R., Defeo, O., 2011. Leadership, social capital and incentives promote successful fisheries. Nature 470, 386–389. Harborne, Alastair R., Mumby, Peter J., 2011. Novel ecosystems: altering fish assemblages in warming waters. Curr. Biol. 21, R822–R824. Haynie, A.C., Pfeiffer, L., 2012. Why economics matters for understanding the effects of climate change on fisheries. ICES J. Mar. Sci. 69, 1160–1167. Held, I.M., Soden, B.J., 2006. Robust responses of the hydrological cycle to global warming. J. Clim. 19, 5686–5699. Hilborn, R., Kennedy, R.B., 1992. Spatial pattern in catch rates: a test of economic theory. Bull. Math. Biol. 54, 263–273. Huitric, M., 2005. Lobster and conch fisheries of Belize: a history of sequential exploitation. Ecol. Soc. 10, 21. Johnson, A.E., Cinner, J.E., Hardt, M.J., Jacquet, J., McClanahan, T.R., Sanchirico, J.N., 2013. Trends, current understanding and future research priorities for artisanal coral reef fisheries research. Fish Fish. 14, 281–292. Jones, R., Noguer, M., Hassell, D., Hudson, S., Jenkins, G., Mitchell, J., 2004. Generating High Resolution Climate Change Scenarios Using PRECIS. Met Office Hadley Centre, Exeter, UK. Kittinger, J.N., Finkbeiner, E.M., Ban, N.C., Broad, K., Carr, M.H., Cinner, J.E., Gelcich, S., Cornwell, M.L., Koehn, J.Z., Basurto, X., Fujita, R., Caldwell, M.R., Crowder, L.B., 2013. Emerging frontiers in social-ecological systems research for sustainability of smallscale fisheries. Current Opinion in Environmental Sustainability 5, 352–357. Lopes, P.F.M., Begossi, A., 2011. Decision-making processes by small-scale fishermen on the southeast coast of Brazil. Fish. Manag. Ecol. 18, 400–410. Markus, T., 2010. Towards sustainable fisheries subsidies: entering a new round of reform under the Common Fisheries Policy. Mar. Policy 34, 1117–1124. Munday, P.L., Jones, G.P., Pratchett, M.S., Williams, A.J., 2008. Climate change and the future for coral reef fishes. Fish Fish. 9, 261–285. Nakicenovic, N., Swart, R., 2000. Special Report on Emissions Scenarios: A Special Report of Working Group III of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, UK. Perez, A., 2009. Fisheries management at the tri-national border between Belize, Guatemala and Honduras. Mar. Policy 33, 195–200. Perry, A.L., Low, P.J., Ellis, J.R., Reynolds, J.D., 2005. Climate change and distribution shifts in marine fishes. Science 308, 1912–1915. Phillips, S.J., Dudík, M., 2008. Modeling of species distributions with Maxent: new extensions and a comprehensive evaluation. Ecography 31, 161–175. Phillips, S.J., Anderson, R.P., Schapire, R.E., 2006. Maximum entropy modeling of species geographic distributions. Ecol. Model. 190, 231–259. Reutter, B.A., Helfer, V., Hirzel, A.H., Vogel, P., 2003. Modelling habitat-suitability using museum collections: an example with three sympatric Apodemus species from the Alps. J. Biogeogr. 30, 581–590. Schorr, D.K., 2005. Artisanal Fishing: Promoting Poverty Reduction and Community Development Through New WTO Rules on Fisheries Subsidies. UNEP p. 34. Simpson, Stephen D., Jennings, S., Johnson, Mark P., Blanchard, Julia L., Schön, P.-J., Sims, David W., Genner, Martin J., 2011. Continental shelf-wide response of a fish assemblage to rapid warming of the sea. Curr. Biol. 21, 1565–1570. Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tignor, M., Miller, H.L., 2007. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, 2007. Cambridge University Press, Cambridge. Sumaila, U.R., Teh, L., Watson, R., Tyedmers, P., Pauly, D., 2008. Fuel price increase, subsidies, overcapacity, and resource sustainability. ICES J. Mar. Sci. 65, 832–840. Suuronen, P., Chopin, F., Glass, C., Løkkeborg, S., Matsushita, Y., Queirolo, D., Rihan, D., 2012. Low impact and fuel efficient fishing — looking beyond the horizon. Fish. Res. 119–120, 135–146. Tett, S.F.B., Stott, P.A., Allen, M.R., Ingram, W.J., Mitchell, J.F.B., 1999. Causes of twentiethcentury temperature change near the Earth's surface. Nature 399, 569–572. Wilen, J.E., Smith, M.D., Lockwood, D., Botsford, L.W., 2002. Avoiding surprises: incorporating fisherman behavior into management models. Bull. Mar. Sci. 70, 553–575. Wisz, M.S., Hijmans, R.J., Li, J., Peterson, A.T., Graham, C.H., Guisan, A., Group, N.P.S.D.W., 2008. Effects of sample size on the performance of species distribution models. Divers. Distrib. 14, 763–773. Zhang, X., Zwiers, F.W., Hegerl, G.C., Lambert, F.H., Gillett, N.P., Solomon, S., Stott, P.A., Nozawa, T., 2007. Detection of human influence on twentieth-century precipitation trends. Nature 448, 461–465.