Al-Abdulrazzak, D., and Pauly, D. Managing fisheries from space: Google Earth improves estimates of distant fish catches. ICES Journal of Marine Science. doi.10.1093/icesjms/fst178.
Many of us have become comfortable with using Google products in everyday life. You may use Google Maps to get the directions of a sushi place you’d like to try out, or use Google street view or Earth to check out your house or to creep on someone else’s house. But did you know that you can use Google geo products to explore the Earth, Moon, Mars, and even dive into the depths of the oceans?
Al-Abdulrazzak and Pauly are scientists from the University of British Columbia who are a part of the Sea Around Us Project and are deeply interested in sustainable fishing. Scientists must have reliable data to help politicians create effective sustainable fishing policies. For a plethora of reasons, reliable field data over a large region can be difficult to obtain. Al-Abdulrazzak and Pauly bypass the issue of unreliable fishery data by using a resource that is freely available to you and me: Google Earth. They explored the potential of using satellite images from Google Earth to estimate fishery yields. In this study, they were interested in fishing weirs of countries within the Persian Gulf.
Fishing weirs have been used all over the world to catch fish for almost 3000 years. They were a popular tool before industrial fishing, which began during the 1950’s. A fishing weir is a structure that can be made of stone, reeds, or wooden posts. It is placed at the edge of a tidal lagoon or at the mouth of a river. There can be many designs (for example step 1 and step 2) but the main purpose of the tool’s design is to allow water to freely flow and to catch fish. The fish become trapped when they try to escape the enclosure while the tide is receding.
Overfishing and environmentally destructive fishing methods threaten the wild fish population. Overtime, fish species can be harvested to the brink of extinction. Sustainable fishing is an important policy issue for members of the United Nations, including those in the Persian Gulf. The nations in the Persian Gulf include Bahrain, Kuwait, Iran, United Arab Emirates (UAE), Iraq, Saudi Arabia, Oman and Qatar. To address the growing concern of over fishing in the Gulf and globally, members of the United Nations (U.N) are asked to submit fishery catch statistics to the U.N. Food and Agriculture Organization (FAO) every year. While the intentions are good, the implementation of sustainable fishing policy is difficult. It is difficult to assess the yield of small-scale fishing operations because of logistical and political issues. Local fisheries implement policy that is driven for development and not for long-term sustainability. Simply put, regulations are weak or not enforced resulting in fishery data that may be unreliable or limited. However, the effects from a lack of regulation are not limited by political boundaries. Despite government efforts to regulate fishing in the Persian Gulf by fishing licenses and spatial restrictions since the 1960’s, most stocks are either fully depleted or overexploited.
Fishing weirs are conspicuous structures when seen from a satellite. Complications arise during conditions of poor visibility. For instance, the image can be of poor resolution or the location may have had foggy conditions at the time the picture was taken. Al-Abdulrazzak and Pauly used Google satellite images from the years 2005 to 2010.
The year 2005 was great for satellite images because they were taken at low tide, when the weirs are most clearly visible. Al-Abdulrazzak and Pauly surveyed the coast of all Gulf countries (inward of the Strait of Hormuz) three times using satellite images from the years 2005 to 2010. They marked each weir with the Google marker tool (Figure 2). They measured the three trap components (wing, yard and pocket) using the Google ruler tool calibrated to tennis courts. Oman was not included in the study. Fishing weirs were found in all counties except Iraq. Satellite images of the Iraq coastline were not clear enough to be used for this study, but the scientists have cited several studies about fishing weirs in Iraq.
In remote sensing studies it is important to “ground-truth” the data. “Ground-truthing” data is to verify the accuracy of the images by comparing them with something of known dimensions or by going out in the field to verify the measurements. Part of the difficulty in this study is that the visibility of weirs depends on the clarity of the image. The image may not be clear during cloudy days. In an ideal situation the scientists would “ground-truth” their data by going to the sites where weirs were not clearly visible in the satellite images. However, this approach is not feasible given the scope of the study.
Al-Abdulrazzak and Pauly addressed this problem by using the images in a unique way. They designed a visibility scheme with five categories (0–20%, 21–40%, 41–60%, 61–80%, 81–100%) based on the proportion of each weir that was visible. A weir that was more visible was closer to 100% while less visible would be farther from 100%. They assigned each weir to one of the categories. They plotted the percentage visible of each weir against the cumulative number of weirs, set the visibility to zero, and solved for the total (seen and unseen) weirs.
Fishing weirs have a mean catch rate. In order to figure out the mean annual catch rate, one can combine the number of fishing weirs to the mean catch rate. Because fishing doesn’t occur year round the total fish caught was corrected by the length of the fishing season. The final fishing weir catch yield was slightly reduced after the correction. The mean catch rates used to calculate total catch were 62.2 kg per day (number of samples = 3; standard deviation = 32.4 kg per day) for Kuwait (values from Kuwait Institute for Scientific Research) and 42.6 kg per day for Bahrain (values from the Department of Fish Resources). Species composition ratios were used to determine how much of each species was caught. Estimates for fishing season lengths and species composition ratios were also obtained from the same sources, as well as for the UAE (Environmental Agency of UAE). Regional averages were applied for countries without these data (Iran, Saudi Arabia and Qatar). FAO catch data are available for weirs in Bahrain, the UAE and Kuwait. Iran, Saudi Arabia or Qatar did not report catch data or number of weirs to the FAO.
The catch rates had a mean and standard deviation. In a perfect world, there would be no error in a measurement. However we are not in a perfect world. We use statistics to tell us how close we are to the true value, despite of the error that is procedural or random. This means that we can say with statistical confidence that the catch rate is within the range of the standard deviation. However, let me ask you this: when we calculate the annual catch rate, will the annual catch rate vary dramatically if we had a scenario with very little error versus a lot of error? Many scientists encounter this question and can answer it using Monte Carlo simulations. Scientists can use different scenarios (for example, a scenario with a lot of error verses a little error) to simulate different conditions. Each calculation is a simulation. A scientist will do many simulations (for example 1000 simulations). In the end they will be able to answer this question: by how much will my calculated value vary under all possible scenarios? This tool is very important for scientists because it allows them to put a number into the many “what if” scenarios.
Al-Abdulrazzak and Pauly located 1656 weirs in the coastal waters of the Persian Gulf. The correction for low visibility yielded an increase of 6.2% in the estimated number of weirs (Figure 3). The correction for imagery availability led to an increase of 8.0%. Jointly, these two factors raised the number of weirs estimated to be in the Persian Gulf in 2005 to 1900 weirs (± 79), with an estimated catch of 31,433 tons (±9827). Compared with regional catches reported by FAO in the same year, these estimated weir catches represent additional catches of 6–8%. The largest weirs were found in in the UAE (321 meters), followed by Iran (222 meters) and Qatar (179 meters).
The majority of the weirs were found in Bahrain (50%), followed by Iran (37%) and Kuwait (5%). Bahrain also leads in estimated catch (54%), followed by Iran (39%). Estimated catches for Iran, Qatar and the UAE represent an additional 4, 2 and 0.01%, respectively as compared with total catches reported to the FAO. Estimated catches for Bahrain were 142% greater than total catches reported to the FAO.
Kuwait’s catch is dominated by mackerel (Scombridae; 30%), bream (Sparidae; 19%) and morraja (Gerreidae; 15%), while Bahrain’s catch is dominated by rabbitfish (Siganidae; 26%), swimming crab (Portunidae; 26%), and sardines (Clupeidae; 15%). Over 96% of the UAE’s catch is reported as snapper (Lutjanidae). The regional average species composition is dominated by Siganidae (22%), Portunidae (22%) and Clupeidae (13%).
Al-Abdulrazzak and Pauly estimate that 1900 weirs found in this study contribute to a catch of 31,433 tons per year. This estimation is six times greater than the FAO reported catch of 5,260 tons.
FAO catch data are available for weirs in Bahrain, the UAE and Kuwait. Their estimates are nine and two times higher than reported for Bahrain and the UAE, respectively, but rather close (within 300 tons) for Kuwait.
No data on catch or number of weirs were reported for Iran, Saudi Arabia or Qatar. Al-Abdulrazzak and Pauly estimate weir catches for Saudi Arabia and Qatar to be less than 500 tons per year, a small number compared to the total annual catch rate. Iran, one the other hand, is estimated to catch over 12,000 tons per year with their 728 weirs. Very little information is available on the Iranian weir fishery even though it is largest in the region and one of the largest fishery regions in the world.
It is clear that the numbers reported to the FAO are fishy. The findings made by Al-Abdulrazzak and Pauly show the potential of satellite imagery and remote sensing to provide more accurate catch numbers and expose illegal fishing practices. It is clear that in order to address the issue of overfishing, scientists must be willing to work with other sources of data that are cost-effective and practical. Google Earth is an excellent place to start.