A mitigation technique for whale-vessel collisions is of interest to both commercial and conservation sectors as they often result in damage to the ship and death of the whale. The scale of which this is happening is difficult to monitor as there is no universal database on the matter and ships are often nervous to report collisions due to perceived implications. However, the IWC Vessel Strike Data Standardisation Committee has been collating both formal and informal reports of whale-vessel strikes since 2005 which indicates whale-vessel strikes are on the rise and unsurprisingly more common in areas with high maritime traffic.
To investigate how to monitor both best, it is essential to look at how they are currently observed separately. However skip forward to the discussion for an overview.
Tracking boats is an established sector that uses a universal technique: AIS (Automatic Identification System). AIS communicates a wide variety of information such as:
- Rate of turn
- Estimated time of arrival
- Ship name
- Type of vessel
The International Maritime Organization (IMO) requires all boats over 300 tonnes to carry AIS. However, uptake of AIS is high across all boat types. The AIS transponders were initially created for local communication between ships and a terrestrial control tower. Therefore, the range is small (<50km2). However recently companies such as exactEarth and ORBCOMM have found a way to increase this range via satellite. This allows monitoring of global traffic, opening up the access to monitor remote regions and repurpose the data. The transponders on the boat transmit every two seconds providing a real-time service.
Satellite AIS works by using a constellation of satellites that orbit the earth. The constellation consists of microsatellites which are 10 to 100kg and low orbiting, taking 90 to 100 minutes to orbit allowing for extensive coverage of the Earth’s surface in frequent intervals. The satellite receives the AIS transmission from every ship in its field of view, which covers around 5000km2. This large footprint allows each satellite in the constellation to receive 4,000,000 messages a day. Satellite coverage of AIS is increasing, and resources are being invested in this supported by the fact Sentinel 1C and 1D are due to carry AIS.
Whale tracking is not as common practice as ship tracking and there is not a universal method in place. Some techniques have been tested using both remote sensing and traditional methods. This post will focus on remote sensing techniques. However, it should be acknowledged that there are other successful techniques such as crowdsourcing sightings and having lookouts on the ship (See: Vanderlaan and Taggart, 2007). I will highlight some of the most popular methods and their benefits and limitations.
Acoustic monitoring uses microphones, called hydrophones, placed on the sea floor on buoys or trailed behind ships. The hydrophone detects changes in pressure from incoming sound waves, which it then converts into electricity. The speed and distance a sound wave travels affect pressure which is reflected in the electrical output. Therefore the location of the whale can be found by the strength of the output. Example
- Ability to make temporal observations
- Can be automated
- Access to remote areas
- The frequency of data collection is high
- Can be swept away by strong currents, collected by fishing nets
- Need previous knowledge of whale vocalisations
- The sound could be a disturbance to the whales
- Only useful in flat and narrow bathymetry
- The exact location is not always computable
- Not appropriate for non-vocal whales (sleeping or solo), which is the most vulnerable time
Habitat monitoring uses satellite imagery to monitor sea surface temperature and chlorophyll which can aid in predicting the presence of whales. Satellites collect electromagnetic radiation reflected from the Earth via a sensor. This works as different surfaces reflect radiation differently. In this case, water is effected by blue and green light differently. Chlorophyll will absorb blue light and reflect green. Therefore through post-processing of the images distinctions can be made. Example
- Can aid in creating ship corridor diversions
- Can assist in deciding locations to put other methods of monitoring, e.g., acoustic buoys
- Contributes to the scientific understanding of whales
- Low cost
- No actual tracking of a whale
- Not all whales have the same habitat
- Very presumptuous
Satellite tracking tags work by placing a transmitter directly on the whale. These send signals in set time periods until machine expires, typically once a day. The tags gather information and transmit when the whale surfaces and connection can be established. Distance travelled is calculated by taking average positions. Example
- Can track long-distance migration routes for shipping corridor adjustments
- Useful in areas with limited human access
- The accuracy of location can be poor
- Can be extended intervals if whale doesn’t surface for a while or is out of range
- Impossible to put on every whale
- Practicalities of placing a tag physically on the whale
Analysis of satellite imagery. It is possible to use high-quality satellite imagery to find whales. To automate the process a filter the size of a whale is applied. The filter will pick out anomalies in the image of that size. These anomalies can then be assessed to see if it a whale, debris, cloud etc. Using the satellite WorldView2, it is possible to view whales in all 9 bands, and the far blue band found whales below the surface. Example
- Can visually see the whales
- Automated processes
- Can make other measurements such as size and type
- Relatively low cost
- Whales dive deep and satellite imagery only penetrates a relatively small section subsurface
- Time-consuming to process image
- Resolution can be poor
- Time differences between photographs can be poor
- Cloud cover can be detrimental
- The whale will have left the area by the time it identified in the imagery
Hyperspectral imagery. A plane with a hyperspectral camera flies 1km over whales, taking 200m wide photographs in 72 spectral bands and 1m2 resolution. GPS then references the data to local navigational charts. How the sensor works: the reflected sunlight enters the sensor through an entrance slit. The light is reflected by a mirror and passed through a dispersing element. The dispersing element diffracts the light into different wavelengths. It is then reflected off another mirror onto a detector array. Post-processing methods are used to remove sea surface reflection. Images are processed as they are for satellite imagery. Example
- Can visually see the whales
- Automated processes
- Can make other measurements such as size and type
- Whales move at the speed so can be tricky to follow them
- Need to know where they will be before them
- Weather conditions need to be ideal
- Is not temporal
- Can’t just fly a plane above all the time.
A unified system
To create an effective and robust mitigation system the two methods of tracking need to be unified. As the AIS system is used extensively, it is logical the whale system be built into this.
There are currently a number of projects to address this bridge. The first is a project called www.listenforwhales.org. The project involves acoustic whale detection in the Stellwagen Sanctuary, an area where collisions are high risk.
As seen in above buoys have been placed in strategic positions throughout the sanctuary. If a buoy detects a whale in the previous 24 hours, ships passing by the buoy are notified and required to reduce their speed to 10 knots.
For the success of this method, the call of the whale has to be known, and the buoy has to be configured to the correct frequency receive this. The main limitation is that whale also has to be vocally active which it is not if it is alone or sleeping, which is are known to increase the risk of a strike. However, another method has been developed which does not have this limitation: masking detection. It is being used by the Canary Islands Whale Anti-Collision Systems (WACS) and is the use of a hydrophone to produce a sound shadow. These shadows highlight the presence of a whale including those which are silent.
http://www.listenforwhales.org is utilising AIS to distribute the alerts from the buoys. The warnings are sent via satellite to the ship which is decoded by software on the ship and overlaid on its AIS system. At present, 1 message for each buoy is sent every 5 minutes. However, there are limitations: due to the poor interoperability of the AIS system, there is no way of knowing if the ship actually received the message. They suggest the introduction of a specific whale AIS channel so that the receipt of information can be acknowledged and ships could send intelligence back.
Continuing from the suggestion of improved communication with ships, there is another project focusing on the ships ability to return their own observations. REPCET (above) is an interface sperate from the AIS system that ships can report whale sightings on. These are transmitted by satellite in real-time to a server located on land, the server then centralises the data and sends out an alert to all vessels in the vicinity. However I believe due to this being outside the AIS uptake is an issue, though the creator’s report good uptake in the study area.
Through consideration of options, it has become apparent that frequency of the collection of data is an important issue. Information needs to be collected at a similar rate to that of the ships as they are both moving entities. Old data is not useful for a ship which has now moved out the area, information needs to be relevant and fast to allow for the vessel to take avoiding measures. To receive this level of frequency, satellite communication is necessary between the whale detector and ship. Furthermore, this data needs to be readily understood by ship captains. Therefore using the satellite imagery methods is not suitable. This is due to the resolution of the data, the dependency of whales being near the surface and the speed at which an image can be processed. On this basis tracking tags placed on whales are also unsuitable. It is impossible to put trackers on every whale, in addition to accuracy and longevity of these systems being questionable. This leaves acoustic monitoring techniques, this technique has the subsurface range, the interoperable format and the potential for accurate real-time monitoring.
Also, when considering a viable mitigation technique, it is important to consider costs. Solutions range from very low cost and low tech methods such as creating ATBAs (Area To Be Avoided) to creating a centre of expertise to monitor this issue globally. To keep costs realistic a collaboration of systems could be proposed. Many systems appear to be developing parallel. Each of the unified methods highlighted above has merits and flaws. Furthermore, they seem to answer each other’s issues, listentowhales.org can only detect vocal whale while the WACS method would combat this. Moreover, AIS has been repurposed for other successful global efforts in the past, showing the potential of harnessing this system. It would be interesting to investigate if a whale monitoring aspect could be integrated into this existing infrastructure such as CATAPULTS illegal fishing monitoring team. WhaleWatch or the Global Fishing Watch.
In conclusion, a method of building whale alerts into the traditional tracking system AIS would be a valid mitigation technique. This would be in the form of a new channel that transmits acoustically collected data on whales and distributes the information in a readable format to nearby vessels. The suggestion of integration into another established program is proposed.
The potential to use data collected by other remote sensing techniques is also notable. Using satellite imagery and tracking tags enrich the data available on whale movement and can aid in the creation of non-earth observation techniques such as adjusting shipping corridors and reducing speed.