Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

On the Potential of Surfers to Monitor Environmental Indicators in the Coastal Zone

  • Robert J. W. Brewin ,

    robr@pml.ac.uk

    Affiliations Plymouth Marine Laboratory, Plymouth, Devon, United Kingdom, National Centre for Earth Observation, Plymouth Marine Laboratory, Plymouth, Devon, United Kingdom

  • Lee de Mora,

    Affiliation Plymouth Marine Laboratory, Plymouth, Devon, United Kingdom

  • Thomas Jackson,

    Affiliation Plymouth Marine Laboratory, Plymouth, Devon, United Kingdom

  • Thomas G. Brewin,

    Affiliation Chatham & Clarendon Grammar School, Ramsgate, Kent, United Kingdom

  • Jamie Shutler

    Affiliation College of Life and Environmental Sciences, University of Exeter, Cornwall Campus, Penryn, Cornwall, United Kingdom

Correction

1 Sep 2016: Brewin RJW, de Mora L, Jackson T, Brewin TG, Shutler J (2016) Correction: On the Potential of Surfers to Monitor Environmental Indicators in the Coastal Zone. PLOS ONE 11(9): e0162591. https://doi.org/10.1371/journal.pone.0162591 View correction

Abstract

The social and economic benefits of the coastal zone make it one of the most treasured environments on our planet. Yet it is vulnerable to increasing anthropogenic pressure and climate change. Coastal management aims to mitigate these pressures while augmenting the socio-economic benefits the coastal region has to offer. However, coastal management is challenged by inadequate sampling of key environmental indicators, partly due to issues relating to cost of data collection. Here, we investigate the use of recreational surfers as platforms to improve sampling coverage of environmental indicators in the coastal zone. We equipped a recreational surfer, based in the south west United Kingdom (UK), with a temperature sensor and Global Positioning System (GPS) device that they used when surfing for a period of one year (85 surfing sessions). The temperature sensor was used to derive estimates of sea-surface temperature (SST), an important environmental indicator, and the GPS device used to provide sample location and to extract information on surfer performance. SST data acquired by the surfer were compared with data from an oceanographic station in the south west UK and with satellite observations. Our results demonstrate: (i) high-quality SST data can be acquired by surfers using low cost sensors; and (ii) GPS data can provide information on surfing performance that may help motivate data collection by surfers. Using recent estimates of the UK surfing population, and frequency of surfer participation, we speculate around 40 million measurements on environmental indicators per year could be acquired at the UK coastline by surfers. This quantity of data is likely to enhance coastal monitoring and aid UK coastal management. Considering surfing is a world-wide sport, our results have global implications and the approach could be expanded to other popular marine recreational activities for coastal monitoring of environmental indicators.

Introduction

The coastal zone is regarded as one of the most valuable and vulnerable habitats on Earth [1]. It contains the richest level of marine biodiversity [2, 3] and has a higher economic value per unit area than terrestrial and open-ocean ecosystems [4, 5]. The coastal zone supports a significant proportion of the world’s fish catch and is a source of non-renewable and renewable energy, waste disposal and recreation [610]. Human population densities in coastal regions (within 100 km distance of the coast and < 100 m above sea level) are estimated to be three times higher than global averages [11] and set to increase [12]. As a consequence the coastal zone is under increasing threat from: overfishing [6]; habitat degradation [13]; marine species loss [14, 15]; climate change [16]; harmful algal blooms [17]; hypoxia [18]; and eutrophication [19]. Coastal management is used to minimise the negative impacts of anthropogenic activity without compromising the socio-economic benefits of the coastal region [20].

Monitoring the coastal zone is fundamental to coastal management. Without adequate monitoring, environmental managers lack the information required to develop sufficient understanding for good management, or enable response to sudden (e.g. sporadic events) and long-term change (e.g. climate change). Environmental indicators are simple measures used to track the state of an environment [21]. In the coastal zone, these indicators can be physical (e.g. changes in land cover, currents, temperature, salinity, turbidity), biological (e.g. phytoplankton abundance and composition, macrophyte abundance) or chemical (e.g. nutrient concentrations, pH, toxic contaminants) [22]. There is high demand for observations on environmental indicators for coastal management of water quality, conservation, human resources and recreation (e.g. European Union Water Framework Directive) [22, 23]. Traditionally in situ measurements, acquired using conventional platforms such as research vessels or buoys, have been used to monitor environmental indicators. However, traditional methods for collecting in situ measurements are expensive and hampered by challenges in the coastal zone; for instance, from biofouling, and from the effects of tides, wave shoaling and coastal currents. The deployment and maintenance of such systems are also inherently expensive. Demand for observations on environmental indicators is not met by in situ datasets currently available [24]. Consequently, inadequate sampling coverage in the coastal zone is regarded as a major challenge facing coastal management [25].

To improve sampling coverage remote-sensing systems have been deployed from satellite, aircraft and at fixed positions along the coastline. Satellite remote-sensing of visible and thermal imagery is capable of providing affordable imagery with good temporal and geographic coverage, but is often limited by spatial resolution [26], and challenged by the optical complexity of coastal waters [27] and by atmospheric-correction [26, 28]. Aircraft-mounted sensors can significantly improve spatial coverage in coastal waters [29, 30], but are costly, especially when acquiring a high-temporal coverage. Fixed video systems are capable of improving spatial coverage [25, 31] and can extract information at low cost, on coastal morphology, currents and waves [3234]. However, fixed video systems are limited in viewing range (∼ 2 km from the cameras in either direction [25]) and only available at specific locations. Furthermore, measurements of the ocean from remote-sensing platforms (satellite, aircraft and from fixed positions along the coastline) are limited to what can be measured using optical and infra-red radiation and require in situ data for calibration and validation. Other innovative solutions are needed to improve sampling coverage in the coastal zone, such as citizen science.

Citizen science is the outsourcing of a task once performed by a set of professionals to a large network of voluntary citizens. If carefully constructed, it can promote public understanding of science [3537] and tackle costly, intractable and laborious research problems [3840]. The generation of reliable scientific data through citizen science has contributed to unexpected insight and innovation, and high-quality research [41, 42]. Emerging technologies, such as mobile applications, wireless sensor networks, on-line computer/video gaming, and miniaturised environmental sensors, show great promise for advancing citizen science [43]. The influence of gaming and competition plays a large role in participant motivation [4345], highlighting the benefits of incorporating recreation into citizen science [46].

In oceanography, citizen science has much untapped potential [47], especially when considering the high cost of oceanographic sampling (e.g. ship or boat hire) in comparison with many terrestrial-based sciences. In the UK alone, it has been estimated that 5.4 million people are involved in a recreational activity that requires direct interaction with the aquatic environment (both ocean and in-land waters), including some: ∼ 800,000 kayakers; ∼ 624,000 small-boat sailors; ∼ 518,000 surfers; ∼ 271,000 scuba divers; and 4.8 million outdoor swimmers [48]. Of all these major water-sport activities, surfing has the highest proportion of activity undertaken at the coastline [48]. The surfing community are also strong advocates of environmental monitoring (e.g. see Surfers Against Sewage [http://www.sas.org.uk/] and the Surfrider Foundation [http://www.surfrider.org/]), orchestrate their activities year round [49], and have an intrinsic interest in the functioning and state of the environment [50], making them a good target audience for citizen science projects in the coastal zone.

In this paper, we investigate the potential of using surfers as platforms to monitor environmental indicators in the coastal zone, with a view to enhance the sampling coverage required to improve coastal management. We focus our efforts on collection of sea-surface temperature (SST) data, considered an important environmental indicator for coastal management [22, 23], which plays a fundamental role in: the structuring of marine biodiversity in coastal environments [3]; the initiation and duration of the spring phytoplankton bloom [51]; the growth and metabolic rates of all trophic level species, from plankton [52, 53] to fish [54, 55]; the exchange of climatically important gases between the atmosphere and the ocean [56]; the local weather and climate [57]. Temperature is also a property that can be measured relatively easily (e.g. through measurements of electric resistance) and cheaply, making it ideal for citizen science-based projects. We equipped as surfer with a temperature sensor and Global Positioning System (GPS) device for a period of one year that they used when surfing. The SST data are compared with estimates from a local oceanographic station and satellite data, to determine if SST acquired from a surfer is reliable and what additional benefits it may bring. The GPS data is used to acquire information on surfer performance and used to illustrate potential motivation for data collection. Our results are extrapolated using estimates of the UK surfing population to demonstrate the potential of using surfers to improve sampling coverage of environmental indicators in the coastal zone. Finally, we discuss the implications of our results for other recreational water-sports and for monitoring environmental indicators not accessible by remote means.

Materials and Methods

Statistical tests

To compare SST data acquired by the surfer with those acquired from other sources, we used the squared Pearson correlation coefficient (r2) and the Root Mean Square Error (Ψ), the latter calculated according to (1) where, X is the variable (SST) and N is the number of samples. The superscripts E and M refer to two independent methods of measuring the same variable (e.g. one from the surfer and one from the satellite). The Ψ can also be partitioned into its precision and accuracy (or bias) components, such that Ψ2 = Δ2+δ2, where the precision component Δ is expressed as (2) and the accuracy (or bias) component as (3) Having a term available that distinguishes systematic (δ) and random (Δ) differences was found to be useful in the context of this study. We also computed the slope (S) and intercept (I) of a linear regression between XE and XM. A slope (S) close to one and an intercept (I) close to zero is an indication that the two estimates of temperature agree well.

Equipment

Following advice from the Faculty Research Ethics Committee for Health and Human Sciences Research at Plymouth University, UK, that ethical approval was not required for our study, a recreational surfer (lead author) was equipped with a UTBI-001 Tidbit V2 Temperature Data Logger and a Garmin etrex 10 GPS (Fig 1).

thumbnail
Fig 1. Equipment used in the study and surfer set-up.

(a) Shows the Tidbit V2 temperature logger attached at mid-point to the surfboard leash. HOBOware software and HOBO USB Optic Base Station (BASE-U-4) were used by the surfer to launch the Tidbit V2 temperature logger prior to each session, and then to upload data post session. (b) Shows the GARMIN extrex 10 GPS, water-resistant Aquapac and waist-bag worn by the surfer. Information at one second intervals on location (latitude and longitude), time, distance, speed and orientation for each surf, were extracted from the GPS device post session. (c) Shows the surfer equipped with the sensors, and (d) shows the surfer collecting data during a session at Wembury beach. Consent to publication was obtained from the participant in this figure.

https://doi.org/10.1371/journal.pone.0127706.g001

The Garmin etrex 10 device was used to extract GPS information. It contains an EGNOS-enabled GPS receiver, has HotFix satellite prediction and can track both GPS and GLONASS satellites simultaneously, allowing it to use 24 more satellites than using GPS alone. During each session, the Garmin etrex 10 device was stored in a water-resistant Aquapac inside a waist-bag worn by the surfer (Fig 1), and set to record GPS data at 1 second intervals. Information on location (latitude and longitude), time, distance, speed and orientation for each surf were extracted from the GPS device post session.

The Tidbit V2 temperature logger was attached, using cable-ties, at mid-point to the leash of the surfboard to ensure continuous contact with seawater when surfing (Fig 1), measuring temperature in the top metre of the water column. The waterproof Tidbit V2 temperature logger has an accuracy of 0.2°C over a range of 0–50°C, a resolution of ∼ 0.02°C at 25°C, a stability of ∼ 0.1°C per year, a response time of 5 minutes in water, and a battery life of ∼ 5 years at a > 1 minute logging interval. Three times during the period of study (May and August 2014 and January 2015), the Tidbit V2 temperature logger was compared with a VWR1620-200 traceable digital thermometer (NIST/ISO calibrated, with an accuracy of 0.05°C at the range of 0 to 100°C and a resolution of 0.001°C) at 1°C intervals from 6 to 25°C using a PolyScience temperature bath. On all three occasions, and over the 6 to 25°C temperature range, the systematic bias (δ) between the Tidbit V2 temperature logger and the VWR1620-200 traceable thermometer was < 0.05°C, lower than the accuracy of the VWR1620-200 traceable thermometer, with a precision (Δ) < 0.025°C and an error (Ψ) < 0.05°C. The slope (S) ranged from 0.997–1.000 and the intercept (I) 0.047–0.084°C for the three tests. Results from the comparison indicate the Tidbit V2 temperature logger performed with high accuracy, with low bias and that its performance was stable over the study period. HOBOware software and HOBO USB Optic Base Station (BASE-U-4) were used by the surfer to launch the Tidbit V2 temperature logger prior to each session, and then to upload data post session. Temperature data were collected at 10 second intervals during each surf.

Study site

The tagged surfer was stationed around the coastline of South West UK (Fig 2a). Between the 5th January 2014 and the 4th January 2015 the surfer orchestrated their recreational activity 85 times at a variety of locations (Fig 2a), with 74% of the surfs (63) conducted at Wembury beach near the city of Plymouth (Fig 2b and 2c) at a near weekly temporal sampling rate. A GPS track, taken on the 13th of September 2014, is shown in Fig 2c and illustrates how the surfer switched on the GPS device (and Tidbit V2 temperature logger) in the car park at Wembury (on land) then walked down to the beach and went surfing, before walking back to the car park and uploading the GPS and temperature data. Speed from the same GPS track is plotted as a function of cumulative distance travelled in Fig 2d, with the spikes in speed indicative of the surfer riding waves. The temperature data for the same session is also plotted as a function of time (Fig 2e) and illustrates a large change in temperature between switching the sensor on at the beginning of each session and entering the water (and exiting the water prior to switching the sensor off) with relatively stable temperature readings during the period the surfer was immersed in seawater.

thumbnail
Fig 2. Study site and sampling locations with an example of GPS and temperature data collected by the surfer.

(a) Shows the locations of the 85 surfing sessions in South West UK conducted during the study, overlain onto a NEODAAS AVHRR SST image taken on the 10th September 2014. (b) Shows a plot of Plymouth and surrounding waters with locations of the surfing sessions near Plymouth and of station L4 in the Western Channel Observatory, with data from the AVHRR SST image (10th September 2014). (c) Shows a plot of Wembury beach in Plymouth, with a GPS track taken by the surfer on the 13th September 2014 overlain onto AVHRR SST estimate at Wembury beach (10th September 2014). (d) Shows speed as a function of cumulative distance travelled for the GPS track taken on the 13th September 2014, with the bumps in speed indicative of the surfer riding waves. (e) Shows a plot of temperature data collected by the surfer during the surf session on the 13th September 2014.

https://doi.org/10.1371/journal.pone.0127706.g002

Temperature data processing

Considering the Tidbit V2 temperature sensor was activated before entering the water, and deactivated after leaving the water, many of the initial and final temperature data were recorded when on land. Furthermore, as the sensor has a response time of up to 5 minutes in sea-water, the time between activation and sea-water equalisation varied among sessions. Therefore, the temperature data required processing to extract SST. While it is possible to use GPS data, together with tidal and shoreline information, to determine when the surfer was in the sea, we developed a GPS-independent SST extraction methodology. This was thought useful for occasions when either the clocks of the GPS and the temperature sensors disagree, or if temperatures were recorded without associated GPS. Fig 3a shows a superposition of all temperature data acquired by the surfer at Wembury beach during the study period. The data were normalised such that the start (surfer entered the water and sensor equalised to water temperature) and end (surfer exited the water and sensor beginning to respond to air temperature) is at the same point on the x-axis for each session. The colour scale of Fig 3a indicates the median of the remaining data (used to compute SST) after the exclusion of erroneous data. The methodology used to determine the start and stop times is described below.

thumbnail
Fig 3. Processing method for the temperature data (denoted in the figure as T) acquired by the surfer.

Figures (a), (d) and (e) were normalised such that the start (surfer entered the water and sensor equalised to water temperature) and end (surfer exited the water and sensor beginning to respond to air temperature) is at the same point on the graph (x-axis) for each session. (a) A superposition of all the temperature data acquired by the surfer during the study period at Wembury beach. (b) A typical temperature data set, acquired on the 13th September 2014, showing the start and stop time of the surf (vertical grey lines), mid-point of data collection (blue line), excluded data (grey shaded areas) and the median of the data collected in the sea (considered as the SST). (c) Standard deviations computed using the processing method for data collected on 13th September 2014, with the corresponding fraction of data used to calculate the standard deviations and the 10% threshold used to exclude data collected when the surfer was on land. (d) Shows the temperature of each session presented as in Fig 3a, but with the computed SST (median of remaining data) subtracted. (e) A superposition of the standard deviations for the data collected at Wembury beach.

https://doi.org/10.1371/journal.pone.0127706.g003

Firstly, we make the assumption that the midpoint of the temperature data for each session occurred while the sensor was in the water. This was visually checked using GPS data and found to occur for every session at Wembury beach. Fig 3b shows an example of temperature data collected on the 13th September with the midpoint shown as a vertical blue line. The temperature data for each session was then divided into two equal halves around the midpoint. For the initial half, every data point was removed sequentially and the standard deviation was calculated incrementally with the last data point representing the standard deviation of the midpoint (zero). For the second half, this procedure was repeated but in reverse. This method produced a list of standard deviations which are plotted (dark blue line) in Fig 3c for data collected on the 13th September, with the fraction of data used to compute each standard deviation plotted as the purple line in Fig 3c. The period for which the surfer measured SST (immersed in sea-water) was then taken to be between the first and last points where the standard deviation was less than 10% of the largest standard deviation (Fig 3c dashed blue line). The cutoff of 10% was chosen based on a visual comparison with the timing of the first and last waves caught by the surfer, as estimated from the GPS data. All temperature measurements before and after these points were excluded (shown in the grey areas of Fig 3a–3e), and the median of the remaining data for each session used to compute SST.

Fig 3d shows the temperature of each session presented as in Fig 3a, but with the computed SST (median of remaining data) subtracted, and Fig 3e shows the standard deviations for the data collected at Wembury beach. These figures demonstrate that the method removes nearly all the erroneous initial and final points. Computing the median of the remaining data as the SST (rather than the mean of the remaining data) also minimises the influence of any erroneous data points that may pass through the processing at the margins of the dataset. The method was checked by visually inspecting the temperature data for each session with the GPS data. During one session, where air temperature and sea temperature were very similar (little variance), the method excluded a significant portion of the data collected in the sea. However, even in this singular case the computed SST using our method was not significantly different (within the accuracy of the sensor) from computing the median of the temperature data using the start (surfer entering the water) and end (surfer exiting the water) derived from the GPS (first and last wave caught).

Our method (Fig 3) is based on the assumptions that: (i) the temperature of the sensor in the sea is relatively stable compared with the variability caused by the transition from air to sea; (ii) that the mid-point of the dataset occurred in the sea; and (iii) duration in the sea is longer than duration out of the water. The method would need to be reviewed for conditions where these assumptions are breached.

GPS data processing

The GPS data processing builds upon the recent work of Barlow et al.[58] to provide performance statistics on surf sessions. The GPS data processing was designed to account for variations in the range of surfing speeds due to differing surfer ability, surfboard types and ocean conditions. During each surf session GPS data were logged at a temporal resolution of one second. Each measurement provided a location, speed and bearing. Initially, a pre-processing filter was applied to remove any anomalous data with speeds greater than 15.3 m s−1, the speed of a swell with a significant wave height of 5 m [59], thought not to have occured during sampling.

Following the high speed filter, the speed at which a surfer could be classed as surfing (wave riding) was determined for each surf. This was done using all data where the surfer was not waiting (speeds greater than 0.5 m s−1[58]). The non-waiting speeds were arranged in ascending order and we implemented a version of the Jerome Friedman’s multivariate adaptive regression splines, to fit a number of hinge functions to the data (Fig 4a). The maximum number of model terms were set to five, to force the model to break the data into three regions. This allowed the identification of two break-points in the range of observed speeds. The first break-point represents the transition from paddling to surfing. Note that this transition is likely to change depending on wave size, surfboard type and surfer performance. We took the first break-point in the hinge function to define the minimum surfing speed for a given surf session. This was conducted independently for each surf session.

thumbnail
Fig 4. GPS processing used to estimate surfer performance statistics.

An example from the 30th August 2014. (a) Shows speed from the GPS plotted as a function of data sorted according to speed (grey crosses). Hinge functions were fitted to the data to partition it into three linear fits (red dashed lines), and identify two break-points. The first break-point (black solid line) in the hinge function was used to define the transition from paddling into surfing (wave riding). (b) Bearing (or direction) data for all surfing data points (where speed is greater than at the break-point in (a)). Red line indicates the average surfing direction during the session and the dark blue lines indicate the region with an angular window of ±90° either side of the mean surfing direction, defined as the permissible surfing directions and used to help control quality of the data. The letter n refers to number of data points measured while travelling above the break-point speed. (c) An example of a categorised wave during the session with speed (and bearing) plotted as a function of time. As the speed increases above the break-point speed (denoted BP) the wave starts (blue dashed line). In this case the wave is classified as ended (dark blue line) when the bearing falls outside the permissible surfing directions (see (b)), with the 8 second window (used to check (in this case) if bearing fell back inside the permissible surfing directions up to 8 seconds after it first fell outside) shown as the grey dashed line. (d) Shows speed as a function of cumulative distance travelled for the GPS track over the entire session, with waves classified in light blue shading. N refers to number of waves during the session and Sc (and the horizontal dashed line) denotes the break-point speed (see (a)) used to define the transition from paddling into surfing during this session.

https://doi.org/10.1371/journal.pone.0127706.g004

In addition to travelling at a required velocity, a surfer should also be travelling in the correct direction (approximately toward shore-line). Analysis of a number of individual sessions revealed occasional anomalous GPS data with speeds well above paddling velocities, headed away from shore, usually following the end of wave. As most of the data with surfing speeds are towards the beach, the bearings for all the data with speeds greater than the minimum surfing speed were averaged to give a mean surfing direction (Fig 4b). An angular window of ±90° either side of the mean surfing direction was defined as permissible surfing directions to allow for left and right rides, and prevent these anomalous GPS data for interfering with the performance statistics.

Following the establishment of these two criteria, the GPS data was analysed in chronological order. Once the minimum surfing speed had been crossed and the surfer was travelling in the right direction they were considered to be riding a wave. As the speed of a surfer may vary during a single wave, with some manoeuvres requiring the surfer to slow (or stall) the board, or to change direction sharply (possibly outside of the permissible surfing directions), the surfer speed was allowed to drop below the threshold velocity and outside permitted direction for up to eight seconds before a wave was considered finished (Fig 4c). Note that this does not result in an eight second addition to the duration of each wave, as each wave was terminated at the last valid point.

Following wave identification, any ride lasting less than four seconds [58] was counted as a failed wave and not included in the riding statistics. Having determined the beginning and end of each wave, the wave statistics and the total session information were computed. An example of performance statistics for a full surfing session is shown in Fig 4d. For the calculation of whole sessions statistics, the start of the session was set at two minutes before the first wave and the end at two minutes after the last wave, acknowledging that there is likely to be variability in this assumption. From the GPS processing a summary was produced for each surf session and the results from these were combined to produce an annual summary (number of sessions, waves caught, total ride time and distance). To put the performance statistics of the surfer derived using the GPS in the context of the wider surfing community, the tagged surfer was rated according to the Hutt et al.[60] surfer skill rating scale, that varies from 1–10. A rating of between 5 and 6 (intermediate level) was assigned to the surfer.

Additional data sources

For comparison with the temperature data measured by the surfer at Wembury beach, SST data were acquired from two other sources: station L4 in the Western Channel Observatory (WCO) and from satellite observations of thermal infra-red radiation. Station L4 is a coastal station located ∼ 12 km south west of Wembury beach (Fig 2b) and forms part of the WCO, an oceanographic time series and marine biodiversity reference site in the Western English Channel [61, 62]. An autonomous buoy is operated at station L4 equipped with a WET Labs Water Quality Monitor (WQM), which incorporates WET Labs’ fluorometer-turbidity and Sea-Bird’s CTD sensors, providing temperature, salinity, depth, dissolved oxygen, chlorophyll fluorescence, turbidity and backscattering data. The WQM records SST at hourly intervals, with an accuracy of 0.002°C at a range of -5 to 35°C, and a resolution of 0.001°C. The buoy at station L4 was brought to shore in November 2013 for maintenance and was ready for redeployed in December 2013. However, due to the large storm events of winter 2013–2014, the buoy was not redeployed until March 2014. SST data were acquired for the period 11th March 2014 to 4th January 2015. Daily median SST were extracted from the time series, and we also extracted SST data at the hour closest in time to that acquired by the surfer at Wembury beach during the study period.

Advanced Very High Resolution Radiometer (AVHRR), daily, ∼ 1 km mapped, SST data were acquired from the NERC Earth Observation Data Acquisition and Analysis Service (NEODAAS [63]). This data were received at Dundee and processed in near-real time at Plymouth Marine Laboratory. An example of an AVHRR image, processed by NEODAAS over the South West UK for 10th of September 2014, is provided in Fig 2a–2c. A daily time series of AVHRR SST data were extracted for station L4 and Wembury beach. This time series represented an average of a box of pixels (3 × 3) centred at station L4 (latitude = 50.25, longitude = -4.2167) and Wembury beach (latitude = 50.3160, longitude = -4.0854). Following standard methods, we used a multi-pixel box to increase the possibility of an in situ measurement (taken either at station L4 or Wembury) being available for comparison with the AVHRR data [64].

Results and Discussion

Sea Surface Temperature (SST) results

SST acquired by the surfer at Wembury beach over the study period is plotted in Fig 5a. In January 2014 (day 5) the SST was around 10°C and dropped to approximately 9.5°C by early March (day 60). Between March and July (day 60–210), the SST increased steadily from 9.5°C to 18°C, with a few sporadic increases observed over the period (Fig 5a). In early August (day 225) SST dropped rapidly from 18°C to 15°C, then rose back to approximately 18°C by late September (day 272). Between October 2014 and January 2015 (day 277 to > 365) there was a steady decline in SST from around 18°C to 11°C.

thumbnail
Fig 5. Sea surface temperature (SST) results from the study.

(a) Shows the time series of SST acquired by the surfer at Wembury beach overlain onto the daily median SST data from station L4 during the study period (N refers to the number of samples). (b) Shows a scatter plot of daily match-ups between SST acquired by the surfer at Wembury beach and SST data from station L4. Bracketed statistics refer to use of hourly match-ups between the two datasets. (c) Shows the time series of SST from satellite (AVHRR) at station L4 overlain onto the daily median SST data from station L4 (buoy) during the study period. (d) Shows a scatter plot of daily match-ups between SST from satellite (AVHRR) at station L4 and SST data from the buoy at station L4. (e) Shows the time series of SST from satellite (AVHRR) at Wembury beach overlain onto SST acquired by the surfer at Wembury beach during the study period. (f) Shows a scatter plot of daily match-ups between SST from satellite (AVHRR) at Wembury beach and SST data acquired by the surfer at Wembury beach. Statistics are denoted as follows: r2 is the squared Pearson correlation coefficient; Ψ is the Root Mean Square Error; Δ is the unbiased Root Mean Square Error; δ is the bias; S and I are the slope and intercept of a linear regression respectively; and N refers to the number of match-ups.

https://doi.org/10.1371/journal.pone.0127706.g005

The time series of SST acquired by the surfer at Wembury beach is overlain onto the daily median SST data from station L4 (Fig 5a). There is good agreement between the two SST estimates, as illustrated by scatter plots between daily match-ups (Fig 5b). Over the seasonal cycle the SST data collected by the surfer at Wembury beach explained 93% of the variance in the SST data at station L4, with an error (Ψ) of 0.78°C and a bias (δ) of 0.31°C. When extracting station L4 SST data at the corresponding hour the surfer was immersed in the water (as opposed to comparing daily median estimates from station L4), statistical results improved further, with the surfer explaining 94% of the variance in the SST data at station L4, with an error (Ψ) of 0.73°C (Fig 5b). The SST data collected by the surfer at Wembury beach is seen to capture the general seasonal cycle (lower SST in the winter, higher SST in the summer) and also abrupt changes, such as the decrease in SST in August (day 225 Fig 5a) seen in the station L4 data. Good agreement between the two datasets illustrate the potential of using surfers to acquire high-quality SST data in the coastal environment.

Whereas the statistical tests confirm that SST from station L4 data agreed well with the SST data collected by the surfer at Wembury beach, there were still some marked differences (Fig 5a). During spring, a few sporadic increases in SST of around 2°C were observed in the Wembury time series, for instance, day 72 and 104 (see Fig 5a). Both these measurements coincide with small wave heights and clear skies (data not shown), with a late morning low tide and the surfing session occurring during mid-high tide in either the afternoon or evening. In both cases, it may be that the exposed inter-tidal land (which has a much lower heat capacity than the ocean) was warmed prior to the rising afternoon tide during which the surf took place, which together with a diurnal increase in water column temperature, possibly resulted in a localised increase in SST during the later period of the day. Sporadic increases in SST close to the coastline are not surprising when considering the multiple factors that may influence coastal temperature (e.g. land-sea temperature exchange and localised run-off from rainfall) that may not be captured by in situ instruments situated further offshore (e.g. at station L4). In some regions, data on differences between coastal and offshore SST may provide useful information regarding freshwater discharge (e.g. close to estuaries), variations in coastal currents, and the location and intensity of tidal mixing fronts [62, 65, 66]. All of which have implications for biological productivity and the transportation of pollutants and contaminants. Monitoring SST in coastal regions, relative to offshore regions, may also benefit marine recreation (e.g. useful information for water bathers).

The time series of satellite-derived SST at station L4 is overlain onto the daily median in situ SST data from the station L4 buoy in Fig 5c, and a scatter plot of the two SST estimates is shown in Fig 5d. At station L4, the satellite data is in very good agreement with the in situ observations, explaining 97% of the variance in the SST data with an error (Ψ) of 0.46°C, a bias (δ) close to zero, a slope (S) close to one, and an intercept (I) close to zero (Fig 5d). By contrast, comparisons of satellite SST and in situ SST (collected by the surfer) at Wembury beach are not so good (Fig 5e and 5f), with the satellite data explaining only 87% of the variance in the in situ SST with a higher error (Ψ = 1.37°C) and a large bias (δ = −0.81°C, Fig 5f). Furthermore, in comparison with 153 available satellite observations at station L4 over the study period, only 60 satellite observations were available at Wembury beach (Fig 5c and 5e). It is common to observe a fewer number of satellite observations in coastal regions as land-sea adjacency complicates the signal received by the satellite sensor and the nature of the aerosol composition at the coastline complicates atmospheric correction [26]. Higher errors (both systematic (δ) and random (Δ)) in satellite-derived SST at Wembury and fewer observations, when compared with those at station L4, further emphasize a need for in situ SST observations at coastal regions, which could be acquired by recreational waters-users such as surfers.

SST is one of the most important characteristics of an aquatic system. It is considered by the Global Climate Observing System as an essential climate variable [67], influencing: dissolved oxygen levels; the solubility and reaction rates of chemicals; the metabolism, growth and reproduction of marine organisms [52, 53]; and water density and stratification, which impact coastal physics and the transport of nutrients, contaminants and pollutants. Variations in nearshore SST have been correlated with coral bleaching events [68] and unusual, and sometimes harmful, algae blooms [69]. Monitoring SST in the nearshore region is of particular importance considering its high level of biological diversity, productivity, and economic value, when compared with open-ocean environments [25], and its vulnerability from a rising human population and climate change.

In addition to SST, it is feasible to monitor other environmental indicators by using recreational water-users, such as surfers. Salinity, a valuable environmental indicator in the coastal environment [70], can be estimated from measurements of conductivity. The acidity (pH) of seawater, which influences metabolic rates and immune responses of some organisms [71] and calcification [72], can also be measured electronically using a pH meter. In fact, as part of the Wendy Schimdt Ocean Health XPRIZE, there are ongoing efforts to develop a surfboard fin capable of simultaneously measuring temperature, salinity, and pH [73]. Monitoring nutrient concentrations and toxic pollutants is important for management of eutrophication and water quality. Recent advances in miniaturised technology [74, 75] may permit future measurements of nutrient concentrations and toxic pollutants from platforms such as surfboards. Large-scale data collection by recreational water-users on environmental indicators has the potential to improve our understanding of the coastal system, that is currently based on observations with suboptimal spatial and temporal coverage. Should a scientific question arise on a component of the coastal system, it may be possible to address this question by equipping recreational water-users with miniature sensors suitable for measuring the environmental indicator pertinent to the question.

GPS results.

Annual surfer performance statistics derived from the GPS processing are provided in Table 1. Over the entire year (85 surf sessions), the surfer caught approximately 2012 waves, surfed for 90.6 hours (nearly 4 days), covered a distance of 375.8 km of which they rode 110.8 km. We conducted a sensitivity analysis of the fixed parameters involved in the GPS processing (see Table 1) and found the GPS processing method relatively robust, with minor effects on the annual statistics (Table 1).

To assess the performance of our GPS method relative to those in the literature, we computed surfer performance metrics for comparison with Barlow et al.[58] (Table 2). In general, our results are remarkably similar with those of Barlow et al.[58] (see Table 2). Some level of similarity can be expected, when considering both methods rely on GPS data and there were similarities in data processing, yet the study of Barlow et al.[58] was based on 60 surfing sessions and 39 recreational surfers of varied ability, whereas our study was based on 85 surfing sessions by the same surfer, of an intermediate level [60]. Statistical results derived here (Table 2) are also comparable with those derived elsewhere [76, 77], especially when considering differences in surfer ability among studies.

thumbnail
Table 2. Surfer performance statistics derived in this study compared with those of Barlow et al. [58].

https://doi.org/10.1371/journal.pone.0127706.t002

As highlighted in previous studies [58, 77], surfing statistics acquired from GPS data have use in monitoring fitness and performance. The prevalence of technology in recreation and leisure has proliferated in recent years with the development of mobile application technology. Mobile phone applications such as Strava, a GPS enabled smart-phone application designed to track recreational activities such as cycling and running, are becoming increasing popular. Coupled with developments in social media (e.g. Twitter and Facebook) and website visualisation (e.g. Google Earth), users of GPS enabled mobile applications can view and share their GPS track and performance. This has led to the use of mobile applications for recreational competition, for instance, the Strava software “King Of The Mountain” that is designed for users to compete with each other over who can cycle a particular route the quickest. The company RipCurl has introduced a GPS surf watch (http://www.ripcurl.com/searchgps-1.html) and other GPS-based devices are becoming available for monitoring surfer performance (e.g. the action sports tracker (TRACE) http://www.traceup.com/). It is likely similar applications will soon be available for surfers to compare their performance and compete. In citizen science, the influence of gaming and competition plays a large role in participant motivation [4345]. Use of mobile GPS technology in surfing for competition could drastically enhance participation in a marine-based citizen science project focused on monitoring environmental indicators like SST.

Potential of surfers to monitor environmental indicators

The equipment and approach used in our study (Fig 1) was designed to test the feasibility of using surfers to monitor environmental indicators in the coastal zone. It would likely require modification for widespread use. For instance, sensor launch, data offload and data post-processing (both SST and GPS) currently require a small investment of time before and after each surfing session, which may discourage citizen uptake. By leveraging emerging citizen-based technologies (e.g. mobile phone applications) and commercial GPS equipment, making use of established methods for data processing, sensor communication and data storage, this could be made quicker, easier and more efficient. As thermistor-based devices become cheaper and more widespread, equipment costs for measuring SST could be reduced further, but would need to balance against data quality and sensor durability. Considering the surfer as a platform, mounted environmental sensors need to be unobtrusive, so as not to interfere with surfer performance and discourage uptake. We found that the Tidbit V2 temperature logger (size = 30 × 41 × 17 mm, weight = 23 g) met this requirement, but not all sensors are likely to. Furthermore, any widespread data collection by surfers is likely to be biased toward conditions and locations preferable for surfing, and subsequent data analysis should consider associated biases. This issue could be minimised if the approach was expanded to other popular marine recreational activities, which take place in a variety of maritime conditions.

Results from this study demonstrate that it is feasible to use surfers as a platform to acquire high-quality data on environmental indicators in the coastal environment (Fig 5). If the approach was scaled-up to a large citizen-science based project, it would be pertinent to gauge the potential in data acquisition by surfers. Knowledge on the global surfing population and the frequency at which surfers partake in their activity is relatively scarce. However, in certain regions, such as in the UK, statistical estimates of surfers are available. For instance, it has been estimated that there are in the region of 500,000 to 700,000 surfers in the UK [48, 78].

Using a survey of > 2000 respondents, and based on an estimate of 500,000 surfers [48], Mills and Cummins [49] estimated the regional distribution of UK surfers and the monthly frequency of participation. Multiplying the number of surfers per region by the annual frequency of participation gives an estimation of the potential number of measurements that could be collected by surfers per year, assuming all UK surfers were equipped to measure environmental indicators. Based on the study of Mills and Cummins [49], we estimate in the region of 40 million independent measurements on environmental indicators could be collected by surfers in the UK per year. Fig 6 shows the spatial (region) and temporal (monthly) distribution of these estimates. South-West England (Devon and Cornwall) is predicted to have the largest benefit with ∼ 18 million measurements per year, followed by the South coast of England, Wales and the East coast of England with between 5–8 million measurements per year each, and by Scotland and Northern Ireland with ∼ 2 million and ∼ 0.4 million measurements per year, respectively. In all six regions, the distribution of measurements are biased toward autumnal months (when the water temperature is warm and the surf conditions relatively consistent in the UK), with a lower number during winter months (Fig 6). As highlighted by Mills and Cummins [49], their study is not without limitations and these numbers are there to serve as rough estimates. Our estimates of sample coverage based on their study (Fig 6) also assume all UK surfers were to participate in data collection when in reality only a small fraction of the total community is likely to participate in a citizen science project. Nevertheless, even if one in a hundred UK surfers were to collect data it is estimated in the region of 400,000 measurements per year in the UK could be collected by surfers on environmental indicators based on study of Mills and Cummins [49]. This would significantly improve sampling coverage in the coastal zone and impact UK coastal management.

thumbnail
Fig 6. Estimates of the number of measurements on environmental indicators that could be acquired by surfers in the UK per month.

These statistics were computed as follows. Firstly, the total number of surfs per month were computed for a series of regions around the UK by multiplying Table 1 of Mills and Cummins [49] (the number of surfers per region) with Table 20 of Mills and Cummins [49] (the number of times per month by region that surfers go surfing). The regions defined in Mills and Cummins [49] were then aggregated into six key areas: South-West England = Cornwall + South Devon + North Devon; South coast England = South Coast 1 + South Coast 2 + South Coast 3; East coast England = East Coast + North East; Scotland = East Coast Scotland + Morray Firth + North Coast + Outer Hebrides + Orkney Islands + Inner Hebrides; Northern Ireland = Northern Ireland; and Wales = Cardiff + Swansea + West Wales + North Wales. Total measurements for each region per year are also provided. Note that land locked UK surfers defined by Mills and Cummins [49], who are also estimated to surf ∼ 2.6 × 106 per year, were not included in this analysis as it was difficult to determine surfing locations.

https://doi.org/10.1371/journal.pone.0127706.g006

Whereas we have illustrated the potential of environmental data collection by surfers in the coastal zone of the UK (Fig 6), surfing is a world-wide activity and conclusions drawn here resonate globally. Surfers frequent uninhabited (e.g. vulnerable reef environments) and remote regions (some of which are inaccessible by large research vessels), and countries with little or no coastal monitoring and coastal management (e.g. parts of Asia, Africa and South America) that may benefit significantly from such data collection. Furthermore, although we have focused our analysis on data collection by surfers, our results have implications for other recreational marine activities. At the UK coastline (excluding in-land waters) it is estimated that there are approximately: 337,000 kayakers; 362,000 small boat sailors; 190,000 scuba divers; 2.8 million outdoor swimmers; 96,000 water-ski and wake-boarders; 98,000 windsurfers; and 62,000 kite-surfers [48]. Citizen science based projects inclusive of these other recreational activities are likely to improve sampling coverage and benefit coastal management.

Conclusions

The coastal zone provides huge social and economic benefits to society yet is under threat from an increasing human population and climate change. Coastal management aims to minimise these negative impacts while maximising societal benefits the coastal zone offers. Unfortunately, coastal management is challenged by inadequate sampling of key environmental indicators. With a view to enhance sampling coverage required for better coastal management, we investigated the possibility of using recreational surfers as platforms for monitoring environmental indicators in the coastal zone. To do this, we equipped a recreational surfer with a GPS device and a temperature sensor for a period of one year. The GPS data were used to extract information on surfer performance, whereas the temperature sensor was used to derive estimates of SST, an important environmental indicator in coastal waters.

By comparing the SST data collected by the surfer at Wembury Beach, UK, with data collected from a nearby oceanographic station (L4) and satellite observations, we conclude that recreational surfers are capable of acquiring high-quality data on SST in the coastal environment. Furthermore, useful information on surfing performance statistics can be acquired from the GPS data that may help motivate data collection by surfers. Based on a recent analysis of the UK surfing population, we estimate that UK surfers have the potential to acquire up to 40 million independent measurements on environmental indicators per year around the UK coastline. Such a huge level of data collection is likely to significantly enhance sampling coverage of environmental indicators required to improve and support coastal management. Surfing is a world-wide recreation, and our results have global implications for: coastal monitoring in remote and under-sampled regions; for other marine recreational activities in the coastal zone; and for monitoring other environmental indicators to that of SST, that may be measured by recreational water-users.

Supporting Information

S1 Data. GPS Tracks.

Contains 85 GPS tracks (comma separated files) used to estimate surfing performance statistics in the manuscript.

https://doi.org/10.1371/journal.pone.0127706.s001

(ZIP)

Acknowledgments

We thank Gavin Tilstone and Vassilis Kitidis for help comparing the temperature sensor (Tidbit V2 temperature logger) used in the study with the VWR1620-200 traceable digital thermometer. We thank Rebecca Jackson for providing photos used in Fig 1 of the paper, and we thank Mark Davidson and Matt Barlow for providing comments on an earlier draft, and Mark Davidson for inspiring the use of GPS to acquire information on surfboard performance. We thank Caroline Hattam for help with inquiries on ethical approval, Pierre Cazenave for discussions on data application and Phyllis Brewin for help proofreading. We also thank the editor, Daniel Harris and one anonymous reviewer for useful comments on the manuscript. The SST data collected by the surfer at Wembury beach and latitude and longitude co-ordinates from the GPS data are publicly available through the British Oceanographic Data Centre (BODC)[79]. The BODC operates in accordance with community standards. We thank Rob Thomas with assistance submitting the dataset to BODC. The GPS tracks for each of the surfs are also available as Supporting Information to this manuscript. The work was supported by Plymouth Marine Laboratory, NERC’s UK National Centre for Earth Observation, and by the Changing Earth Science Network initiative of the STSE program of the European Space Agency. We thank the NERC Earth Observation Data Acquisition and Analysis Service (NEODAAS) for the AVHRR imagery and Ben Taylor for assistance in data processing.

Author Contributions

Conceived and designed the experiments: RJWB LdM TJ TGB JS. Performed the experiments: RJWB LdM TJ TGB. Analyzed the data: RJWB LdM TJ. Contributed reagents/materials/analysis tools: RJWB LdM TJ TGB. Wrote the paper: RJWB LdM TJ TGB JS.

References

  1. 1. Jickells TD. Nutrient Biogeochemistry of the Coastal Zone. Science. 1998;281:217–222. pmid:9660744
  2. 2. Norse EA. Global marine biological diversity: A strategy for building conservation into decision making. Island Press; 1993.
  3. 3. Tittensor DP, Mora C, Jetz W, Lotze HK, Ricard D, Berghe EV, et al. Global patterns and predictors of marine biodiversity across taxa. Nature. 2010;466:1098–1103. pmid:20668450
  4. 4. Costanza R, d’Arge R, de Groot R, Farberk S, Grasso M, Hannon B, et al. The value of the world’s ecosystem services and natural capital. Nature. 1997;387:253–260.
  5. 5. Costanza R, de Groot R, Sutton P, van der Ploeg S, Anderson SJ, Kubiszewski I, et al. Changes in the global value of ecosystem services. Global Environmental Change. 2014;26:152–158.
  6. 6. Stewart KR, Lewison RL, Dunn DC, Borkland RH, Kelez S, Halpin PN, et al. Characterizing Fishing Effort and Spatial Extent of Coastal Fisheries. PLoS ONE. 2010;5:e14451. pmid:21206903
  7. 7. Roc T, Conley DC, Greaves D. Methodology for tidal turbine representation in ocean circulation model. Renewable Energy. 2013;51:448–464.
  8. 8. Monk K, Zou Q, Conley DC. An approximate solution for the wave energy shadow in the lee of an array of overtopping type wave energy converters. Coastal Engineering. 2013;73:115–132.
  9. 9. Cicin-Sain B, Knecht RW, Jang D, Fisk GW. Integrated coastal and ocean management: concepts and practices. Island Press; 1998.
  10. 10. Gill AB. Offshore renewable energy: ecological implications of generating electricity in the coastal zone. Journal of Applied Ecology. 2005;42:605–615.
  11. 11. Small C, Nicholls RJ. A Global Analysis of Human Settlement in Coastal Zones. Journal of Coastal Research. 2003;19:584–599.
  12. 12. McGranahan G, Balk D, Anderson B. The rising tide: assessing the risks of climate change and human settlements in low elevation coastal zones. Environment and Urbanization. 2007;19:17–37.
  13. 13. Lotze HK, Lenihan HS, Bourque BJ, Bradbury RH, Cooke RG, Kay MC, et al. Depletion, Degradation, and Recovery Potential of Estuaries and Coastal Seas. Science. 2006;312:1806–1809. pmid:16794081
  14. 14. Roberts CM, Hawkins JP. Extinction risk in the sea. Trends in Ecology and Evolution. 1999;14:241–246. pmid:10354629
  15. 15. Kappel CV. Losing pieces of the puzzle: threats to marine, estuarine, and diadromous species. Frontiers in Ecology and the Environment. 2005;3:275–282.
  16. 16. Barbier EB. A global strategy for protecting vulnerable coastal populations. Science. 2014;345:1250–1251. pmid:25214595
  17. 17. Anderson DM, Glibert PM, Burkholder JM. Harmful algal blooms and eutrophication: Nutrient sources, composition, and consequences. Estuaries. 2002;24:704–726.
  18. 18. Diaz RJ, Rosenberg R. Spreading Dead Zones and Consequences for Marine Ecosystems. Science. 2008;321:926–929. pmid:18703733
  19. 19. Cloern JE. Our evolving conceptual model of the coastal eutrophication problem. Marine Ecological Progress Series. 2001;210:223–253.
  20. 20. Nobre AM. Scientific approaches to address challenges in coastal management. Marine Ecological Progress Series. 2011;434:279–289.
  21. 21. Platt T, Sathyendranath S. Ecological indicators for the pelagic zone of the ocean from remote sensing. Remote Sensing Environment. 2008;112:3426–3436.
  22. 22. Ferreira JG, Vale C, Soares CV, Salas F, Stacey PE, Bricker SB, et al. Monitoring of coastal and transitional waters under the E.U. Water Framework Directive. Environmental Monitoring and Assessment. 2007;135:195–216. pmid:17342432
  23. 23. Borja A. The European water framework directive: A challenge for nearshore, coastal and continental shelf research. Continental Shelf Research. 2005;25:1768–1783.
  24. 24. Defeo A, McLachlan A, Schoeman DS, Schlacher TA, Dugan J, Jones A, et al. Threats to sandy beach ecosystems: A review. Estuarine, Coastal and Shelf Sciences. 2009;81:1–12.
  25. 25. Davidson M, Koningsveld MV, de Kruif A, Rawson J, Holman R, Lamberti A, et al. The Coast View project: Developing video-derived Coastal State Indicators in support of coastal zone management. Coastal Engineering. 2007;54:463–475.
  26. 26. Thomas A, Byrne D, Weatherbee R. Coastal sea surface temperature variability from Landsat infrared data. Remote Sensing of Environment. 2002;81:262–272.
  27. 27. IOCCG. Remote Sensing of Ocean Colour in Coastal, and Other Optically Complex waters. Sathyendranath S. (e.d.), Reports of the International Ocean-Colour Coordinating Group, No. 3, IOCCG, Dartmouth, Canada; 2000.
  28. 28. Hu C, Carder KL, Muller-Karger FE. Atmospheric Correction of SeaWiFS Imagery over Turbid Coastal Waters: A Practical Method. Remote Sensing of Environment. 2000;74:195–206.
  29. 29. Irish JL, White TE. Coastal engineering applications of high-resolution lidar bathymetry. Coastal Engineering. 1998;35:47–71.
  30. 30. Davis CO, Bowles J, Leathers RA, Korwan D, Downes TV, Snyder WA, et al. Ocean PHILLS hyperspectral imager: design, characterization, and calibration. Optics Express. 2002;10:210–221. pmid:19424352
  31. 31. Holman RA, Sallenger AH Jr, Lippman TC, Haines JW. The application of video image processing to the study of nearshore processes. Oceanography. 1993;6:7885.
  32. 32. Lippmann TC, Holman RA. Quantification of sand bar morphology: a video technique based on wave dissipation. Journal of Geophysical Research. 1989;94:995–1011.
  33. 33. Chickadel CC, Holman RA, Freilich MF. An optical technique for the measurement of longshore currents. Journal of Geophysical Research. 2003;108:3364.
  34. 34. Aarninkhof SGJ, Turner IL, Dronkers TDT, Caljouw M, Nipius L. A video-based technique for mapping intertidal beach bathymetry. Coastal Engineering. 2003;49:275–289.
  35. 35. Trumbull DJ, Bonney R, Bascom D, Carbral A. Thinking Scientifically during Participation in a Citizen-Science Project. Science Education. 2000;84:265–275.
  36. 36. Brossard D, Lewenstein B, Bonney R. Scientific knowledge and attitude change: The impact of a citizen science project. International Journal of Science Education. 2005;9:1099–1121.
  37. 37. Wiggins A, Crowston K. From Conservation to Crowdsourcing: A Typology of Citizen Science. In: Proceedings of the 2011 44th Hawaii International Conference on System. IEEE Computer Society Washington, DC, USA; 2011. p. 1–10.
  38. 38. Gura T. Citizen science: Amateur experts. Nature. 2013;496:256–261.
  39. 39. Schwartz MD, Betancourt JL, Weltzin JF. From Caprio’s lilacs to the USA National Phenology Network. Frontiers in Ecology and the Environment. 2012;20:324–327.
  40. 40. Dickinson JL, Shirk J, Bonter D, Bonney R, Crain RL, Martin J, et al. The current state of citizen science as a tool for ecological research and public engagement. Frontiers in Ecology and the Environment. 2012;10:291–297.
  41. 41. Halberda J, Ly R, Wilmer JB, Naiman DQ, Germine L. Number sense across the lifespan as revealed by a massive Internet-based sample. Proceedings of the National Academy of Sciences of the United States of America. 2012;109(28):11116–11120. pmid:22733748
  42. 42. Jeong SJ, Medvigy D, Shevliakova E, Malyshev S. Predicting changes in temperate forest budburst using continental-scale observations and models. Geophysical Research Letters. 2013;40:359–364.
  43. 43. Newman G, Wiggins A, Crall A, Graham E, Newman S, Crowston K. The future of citizen science: emerging technologies and shifting paradigms. Frontiers in Ecology and the Environment. 2012;10:298–304.
  44. 44. Cooper S, Khatib F, Treuille A, Barbero J, Lee J, Beenen M, et al. Predicting protein structures with a multiplayer online game. Nature. 2010;466:756–760. pmid:20686574
  45. 45. Graham E, Henderson S, Schloss A. Brief report: using mobile phones to engage citizen scientists in geosciences research. Eos, Transactions, American Geophysical Union. 2011;92:313–315.
  46. 46. Nov O, Arazy O, Anderson D. Dusting for science: motivation and participation of digital citizen science volunteers. In: Proceedings of the 2011 iConference (iConference 11); 8–11 Feb 2011; Seattle, WA. New York, NY: Association for Computing Machinery; 2011.
  47. 47. Lauro FM, Senstius SJ, Cullen J, Neches R, Jensen RM, Brown MV, et al. The Common Oceanog-rapher: Crowdsourcing the Collection of Oceanographic Data. PLoS Biology. 2014;12(9):e1001947. pmid:25203659
  48. 48. DEFRA. UK Department for Environment Food & Rural Affairs (DEFRA) “Charting Progress: Leisure and Recreation”; 2007. (accessed 15/12/2014). Available from: http://chartingprogress.defra.gov.uk/feeder/Scetion_3.6_Leisure_and_Recreation.pdf.
  49. 49. Mills B, Cummins A. The economic impact of domestic surfing on the United Kingdom. Wheal Kitty Workshops, St. Agnes, Cornwall, TR5 0RD.: Surfers Against Sewage; 2013. Available from: http://www.sas.org.uk/wp-content/uploads/SAS-Economic-Impact-of-domestic-surfing-on-the-UK-med-2.pdf. pmid:24394589
  50. 50. Butt T, Russell PE, Grigg R. Surf Science: An Introduction to Waves for Surfing. University of Hawaii Press, USA, and Alison Hodge publishers, UK ISBN 0 8248 2891 7 (USA), ISBN 0 906720 36 2 (UK); 2004.
  51. 51. Smyth TJ, Allen I, Atkinson A, Bruun J, Harmer R, Pingree RD, et al. Ocean Net Heat Flux Influences Seasonal to Interannual Patterns of Plankton Abundance. PLoS ONE. 2014;9:e98709. pmid:24918906
  52. 52. Epply RW. Temperature and phytoplankton growth in the sea. Fishery Bulletin. 1972;70:1063–1085.
  53. 53. Keller AA, Oviatt CA, Walker HA, Hawk JD. Predicted impacts of elevated temperature on the magnitude of the winter-spring phytoplankton bloom in temperate coastal waters: A mesocosm study. Limnology and Oceanography. 1999;44:344–356.
  54. 54. Pepin P. Effect of Temperature and Size on Development, Mortality, and Survival Rates of the Pelagic Early Life History Stages of Marine Fish. Canadian Journal of Fisheries and Aquatic Sciences. 1991;48:503–518.
  55. 55. Handeland SO, Imsland AK, Stefansson SO. The effect of temperature and fish size on growth, feed intake, food conversion efficiency and stomach evacuation rate of Atlantic salmon post-smolts. Aquaculture. 2008;283:36–42.
  56. 56. Land PE, Shutler JD, Cowling RD, Woolf DK, Walker P, Findlay HS, et al. Climate change impacts on seaair fluxes of CO2 in three Arctic seas: a sensitivity study using Earth observation. Biogeosciences. 2013;10:8109–8128.
  57. 57. Newell RE. Climate and the Ocean. American Scientist. 1979;67:405–416.
  58. 58. Barlow MJ, Grestry K, Findlay M, Cooke CB, Davidson M. The Effect of Wave Conditions and Surfer Ability on Performance and the Physiological Response of Recreational Surfers. Journal of Strength & Conditioning Research. 2014;28:2946–2953.
  59. 59. Dally WR. The Maximum Speed of Surfers. Journal of Coastal Research. 2001;29:33–40.
  60. 60. Hutt JA, Black KP, Mead ST. Classification of Surf Breaks in Relation to Surfing Skill. Journal of Coastal Research. 2001;29:66–81.
  61. 61. Harris R. The L4 time-series: the first 20 years. Journal of Plankton Research. 2010;32:577–583.
  62. 62. Smyth TJ, Fishwick JR, Al-Moosawi L, Cummings D, Harris C, Kitidis V, et al. A broad spatio-temporal view of the Western English Channel observatory. Journal of Plankton Research. 2010;32:585–601.
  63. 63. Groom S, Martinez-Vicente V, Fishwick JR, Tilstone G, Moore G, Smyth T, et al. The Western English Channel observatory: Optical characteristics of station L4. Journal of Marine Systems. 2009;77:278–295.
  64. 64. Bailey SW, Werdell PJ. A multi-sensor approach for the on-orbit validation of ocean color satellite data products. Remote Sensing of Environment. 2006;102:12–23.
  65. 65. Pingree RD, Pugh PR, Holligan PM, Forster GR. Summer phytoplankton blooms and red tides along tidal fronts in approaches to English-Channel. Nature. 1975;258:672–677.
  66. 66. Simpson JH. The shelf-sea fronts -implications of their existence and behavior. Philosophical Transactions of the Royal Society A. 1981;302:531–546.
  67. 67. Mason PJ, Zillman JW, Simmons A, Lindstrom EJ, Harrison DE, Dolman H, et al. Implementation Plan for the Global Observing System for Climate in Support of the UNFCCC (2010 Update). Geneva, Switzerland, WMO, IOC, UNEP, ICSU, 180pp. (WMO-TD, 1523); 2010.
  68. 68. Goreau TJ, Hayes RL. Coral bleaching and ocean “hot spots”. Ambio. 1994;23:176–180.
  69. 69. Paerl HW, Huisman J. Blooms like it hot. Science. 2008;320:57–58. pmid:18388279
  70. 70. Morris AW, Mantoura RFC, Bale AJ, Howland RJM. Very low salinity regions of estuaries: important sites for chemical and biological reactions. Nature. 1978;274:678–680.
  71. 71. Bibby R, Widdicombe S, Parry H, Spicer J, Pipe R. Effects of ocean acidification on the immune response of the blue mussel Mytilus edulis. Aquatic Biology. 2008;2:67–74.
  72. 72. Orr J, Fabry VJ, Aumont O, Bopp L, Doney SC, Feely RA, et al. Anthropogenic ocean acidification over the twenty-first century and its impact on calcifying organisms. Nature. 2005;437:681–686. pmid:16193043
  73. 73. Lapowsky I. The Next Big Thing You Missed: Surfboard Collects Oceanic Data While You Ride Waves; 2014. (accessed 12/12/2014). Available from: http://www.wired.com/2014/11/board-formula/.
  74. 74. Beaton AD, Sieben VJ, Floquet CFA, Waugh EM, Abi Kaed Bey S, Ogilvie IRG, et al. An automated microfluidic colourimetric sensor applied in situ to determine nitrite concentration. Sensors and Actuators B: Chemical. 2011;156:1009–1014.
  75. 75. Lafleur JP, Rackov AA, McAuley S, Salin ED. Miniaturised centrifugal solid phase extraction platforms for in-field sampling, pre-concentration and spectrometric detection of organic pollutants in aqueous samples. Talanta. 2010;81:722–726. pmid:20188988
  76. 76. Médndez-Villanueva A, Bishop D, Hamer P. Activity profile of world-class professional surfers during competition: a case study. Journal of Strength & Conditioning Research. 2006;20:477–482.
  77. 77. Farley ORL, Harris NK, Kilding AE. Physiological demands of competitive surfing. Journal of Strength & Conditioning Research. 2012;26:1887–1896.
  78. 78. British-Marine-Federation, Maritime-Coastal-Agency, Royal-Yatching-Association, Lifeboats, British-Canoe-Union. Water sports and Leisure Participation Survey 2012; 2012. (accessed on 08/01/2015). Available from: http://www.bcu.org.uk/files/Watersports%20Participation%20Survey%202012%20-%20Executive%20Summary.pdf.
  79. 79. Brewin RJW, de Mora L, Jackson T, Brewin TG, Shutler J. Annual time series of sea surface temperature (SST) measurements collected by a surfer at Wembury Beach, Plymouth, UK. British Oceanographic Data Centre - Natural Environment Research Council, UK. 2015.