The Trouble with Fishing Maps & How to Use Them Better

June 25, 2024

Figure 1. Eddying off the Mahia Peninsula, New Zealand as shown by the distribution of chlorophyll a in the surface waters. Dark grey indicates land. Lines are depth contours in metres. Mass concentration of chlorophyll is plotted on a log10 scale in units of mg/m³. Map coded in Python by the author.

Fishing maps are loaded with promise and fraught with disappointment. There are good reasons for this as I’ll explain in this article. Fishing maps that indicate where fish can be caught have been around for a long time, but the use of satellite remote sensing and hydrographic model data to produce fishing maps is much more recent. The availability of data for download from NASA, NOAA, and the European Space Agency provides companies and individuals with sea surface temperatures (SST), chlorophyll, sea surface height, and current data, among many other variables. Maps based on these data are sold to recreational game fishers to help them find large pelagic fish like marlin, sailfish, swordfish, tunas, and dolphin fish. In some cases, the providers download maps directly from graphical user interfaces, but other companies employ data scientists, remote sensing specialists, and oceanographers to process the data themselves, applying cloud removal techniques and compositing images. The price for these products varies, depending on the sophistication of the fishing maps, from a couple of hundred U.S. dollars per year to thousands of dollars per year for some products used by commercial fishing companies. I produce fishing maps for deep-sea fishing as a product from our small company in Huia, New Zealand (www.fishingmaps.info and https://fishingmaps.company.site).

The production of fishing maps raises some interesting issues involving oceanography, data science, programming, and client expectations. In this article, I focus on (1) the question of scale, (2) why the ocean currents presented on fishing maps are not especially useful, and (3) why the fronts on some fishing maps may not be visible out on the ocean. These issues can be a source of disappointment with fishing maps, but when they are understood, they can help fishers to use the maps more effectively to find the big game fish.

The question of scale

Oceanographers generally work at larger scales than game fishers are interested in. The way oceanographers look at the ocean is very similar to the way meteorologists look at the weather. Both are concerned with the large scale dynamics of heat, pressure, and density, the distribution of which drive the winds and ocean currents. In contrast, game fishers are focused on smaller scales defined by the cost of fuel for their boat, or the boundaries of fishing competitions. A range of 100 km from port would encompass most day fishing trips, which is not surprising considering that 100 km is a 5 hour boat ride at 20 knots. To understand the drivers creating the patterns in a fishing map, an oceanographer will examine the ocean and atmosphere over larger scales (e.g. Figures 2 & 4), perhaps 500 km rather than 100 km (Figure 3). The fishing map displays a subset of much larger patterns, and it can be difficult to determine what is causing the features observed on the small map, unless the larger scale is observed and measured.

Figure 2. An oceanographer will prefer a larger scale map in the first instance because it shows the larger scale structure and dynamics. See also Figure 4. Map coded in Python by the author.

Figure 3. Fishers prefer many smaller maps that reflect the distance that they range from fishing ports. Maps coded in Python by the author.

Current satellite altimetry sensors cannot resolve features smaller than about 50 km although the new SWOT (Surface Water and Ocean Topography) instrument coming on stream in late 2023 will resolve features as small as 15 km in diameter. Large scale patterns of sea surface height cannot be seen in small fishing maps, so the eddies that are ubiquitous in the ocean are poorly resolved. At the larger scale, for example, the development of eddies in the eastern Bay of Plenty, New Zealand can be clearly seen in the chlorophyll a imagery. These eddies propagate along the coast and inject higher production waters into the offshore waters to the east of Hicks Bay. This mesoscale activity cannot be seen as clearly in the smaller fishing maps for Cape Runaway or Hicks Bay.

Figure 4. East Cape of the North Island of New Zealand showing the formation of eddies in the eastern Bay of Plenty and their propagation to the east, past Cape Runaway and Hicks Bay (near the tip of the Cape). The eddies inject higher production into the offshore waters to the east of the cape. Map is from the E.U. Copernicus Marine Service MyOcean Pro viewer.

Eddies can concentrate plankton at their edges or centres, depending on the type of eddy and their direction of rotation. These eddies are one of the important features that game fish are associated with. Some fishing map providers show sea surface height isolines and geostrophic current vectors (shown as arrows) on their maps but they don’t resolve the eddies. It is analogous to trying to interpret the weather map when presented with a partial frontal line, but without being able to see the high and low pressure areas that are evident at larger scales. A take home point is that, at the scale of fishing maps relevant to fishers (100 km or less), sea surface height and current vectors derived from satellite altimetry do not provide useful information about the eddies and fronts where gamefish can often be found. I delve into this more in the next section.

Why are the currents on fishing maps not very useful

In a nutshell, the currents produced by hydrographic models that also use altimetry (among other things) are at the wrong scale for fishing maps. This is because the currents that are associated with aggregations of baitfish occur at much smaller scales. The good thing is that the patterns of SST and chlorophyll seen on the fishing maps are partly due to water movement, so you effectively have at least the the surface currents from the distribution of SST and chlorophyll. A time series of maps over several days shows how these properties are changing.

The currents derived from altimetry and hydrographic model calculations that are often presented on fishing maps are based on large-scale pressure differences. Oceanographers call these geostrophic currents because they maintain the geostrophic pressure balance between waters with different densities. Some models incorporate wind driven (Ekman) currents, and refine the resolution of the product with data from surface drifters, and data from other instruments on moorings. These models show the currents that constitute the river-like flows in the ocean (Figure 5 & 6), but they are not the finer-scale currents associated with frontal regions where baitfish concentrations are often found.

Figure 5. At the large scale, patterns of sea surface height can be seen (highs in red, lows in blue) and these sea surface height differences drive geostrophic current flows, shown by the arrows. There are many eddies evident in the flow field and the flow patterns are coherent. Map is from the E.U. Copernicus Marine Service MyOcean Pro viewer.

Figure 6. At smaller scales the patterns in sea surface height are completely lost, and the currents cannot be estimated over the shelf areas. To do this properly would require a regional ocean model, and near real time data from those models are not currently publicly available in New Zealand for use in fishing maps . Map is from the E.U. Copernicus Marine Service MyOcean Pro viewer.

Baitfish are prey for large pelagic fishes that fishers target, and the baitfish themselves feed on phytoplankton in the case of sardines or pilchards, or zooplankton in the case of anchovies. The conventional idea is that the plankton are concentrated by currents associated with density fronts. These fronts, which often can be seen as lines in the sea, are areas where waters of different density meet and flow over and under each other as they mix. The different densities are caused by differences in temperature and salinity, and may be associated with different levels of nutrients. The phytoplankton grow faster with more nutrients, so the waters may be different colour due to the phytoplankton, or due to sediment content. Fishers will look for the boundaries between more productive green water rich with phytoplankton, and the clear blue water on the other side of a front. Where the waters of different density meet, they form a density front where the denser water sinks beneath the less dense water. This causes convergent flows as the water layers meet and mix, and the convergence accumulates plankton that may swim upward against the sinking current. This combination of behaviour and small scale flows at density fronts creates slicks which may be visible as lines of smooth water, sometimes with accumulations of foam, flotsam, and plankton. These areas of higher plankton density are attractive feeding areas for baitfishes¹.

Phytoplankton grow rapidly, with a turnover time of a couple of days. However, recent research² showed that the residence time of water in a front (how long water stays in a frontal zone) is shorter than the turnover time of the phytoplankton (how fast the phytoplankton population doubles). What this means is that phytoplankton don’t accumulate in fronts by growing there, but instead are washed into frontal zones from more productive areas upstream of the front. The phytoplankton grow in areas where pulses of upwelling occur, and the resulting patches of green water flow into frontal zones. The fronts are conduits, not incubators.

What does all this mean for fishing maps? The small-scale currents in the fronts are not well resolved by the hydrographic models producing the geostrophic currents that are often presented on fishing maps (Figure 6). But since the fronts are conduits,rather than incubators, the important features to map are the phytoplankton patches growing in cooler upwelled water and their subsequent distribution. This is given by the chlorophyll and SST data. The new knowledge of how fronts work means that simpler fishing maps of SST and chlorophyll are going to be sufficient, and adding the large scale geostrophic currents using the data available in 2023 doesn’t add much value. The situation will change soon with the advent of the SWOT (Surface Water and Ocean Topography) sensor data coming online in late 2023, but that’s another story.

Why the fronts on the fishing maps may not be visible out on the ocean

Colours can be deceiving. By using an inappropriate colour scheme for fishing maps, it is quite possible to overemphasize a feature on the map that is not so evident in the data. On fishing maps, the goal is to find the boundaries between green and blue water on chlorophyll maps, and warm and cool water on SST maps. These boundaries occur at density fronts. Fishers sometimes call these features “edges”, but an oceanographer would refer to them as high gradient regions or fronts. Using colour on the fishing map helps to highlight the edges. The most familiar colour scheme for fishers is the “rainbow” colour map, which is similar to the “jet” or “turbo” colour maps. The rainbow colour map does not have a progressive relationship between category and colour. This is not a problem for fishing maps where the emphasis is on quantitative differences in temperature or chlorophyll, rather than on category. However, this colour map also has serious problems that arise from the fact that the colours have differing illuminance. The relationship between change in colour and change in data value is not linear. Yellow has higher illuminance and draws more attention than other colours. The human eye reacts to both colour and illuminance, so the non-linear relationship between colour progression and data increase misleads the viewer to see edges on the map as being stronger than the data justify ³ ⁴. Using the conventional rainbow colour map can mislead the viewer to see sharp fronts that are not strong gradients in the temperature or chlorophyll data.

A specific example helps to elaborate this point. In Figure 7, note how the eddying filaments near 39° 50′ S 178° 20′ E to the east of the Mahia Peninsula, New Zealand are exaggerated by the contrast between red and yellow on the rainbow color map. The colour gradients around these features are much more subtle in Figure 8 which uses the equal intensity viridis colour map. Instead, in Figure 8, the eye is drawn to a hammerhead-shaped intrusion of higher chlorophyll water to the west of the peninsula, and to almost linear frontal features defined by low chlorophyll water to the south east of the peninsula (Figure 8). The different colour maps emphasize different features. My point is that Figure 8 will be a better guide to the frontal zones because the uneven illuminance of the rainbow colour map in Figure 7 misleads the viewer.

Figure 7: Chlorophyll a concentrations for Hawkes Bay, New Zealand, over four consecutive days using the turbo colourmap. Dark grey areas are land. Light grey areas contain no data due to cloud coverage. Chlorophyll data are plotted on a log10 scale. Lines indicate depth contours and colours show the mass concentration of chlorophyll a in mg/m³. Map coded in Python by the author.

Figure 8: These are the same maps as Figure 7 but plotted using the equal intensity viridis colour map. Map coded in Python by the author.

There are several solutions to the color map problem that can be applied to make better fishing maps. The first is to avoid the familiar rainbow colour scheme. There are many perceptually uniform colour maps, but these are unfamiliar to fishers, and consequently they may not want to use maps with those colour schemes. A second improvement would be to use a rainbow colour map that has had the illuminance adjusted to produce a perceptually uniform map. This is what I do in the SST maps that I use in my fishing maps. My SST maps use an illuminance adjusted rainbow colourmap, and my chlorophyll maps avoid the rainbow map, and use the viridis colourmap instead.

A third solution is to calculate the gradients in SST and chlorophyll, select the gradients above some threshold value, and plot those selected values as a line to show where the fronts are. Ideally, the frontal line would be plotted on an perceptually uniform colour map.

Conclusions

Fishing maps are an applied oceanographic product, and like most applications, they are a compromise between the science and the expectations of the clients. The colourful and familiar rainbow colour map can be modified to be more perceptually uniform, yet still remain familiar. However, it is preferable to avoid the rainbow colour map because it is misleading.
More variables, like sea surface height and geostrophic currents, do not add value when they are at the wrong scale and do not resolve target features like mesoscale eddies and fronts. The key variables of SST and chlorophyll plotted over bathymetry can yield enough information to find the places in the ocean where baitfish come to feed, and large pelagic fishes come to eat them. Until data from the new Surface Water and Ocean Topography instrument becomes available in 2024, it is best to use simpler maps that do not mislead with inappropriate variables.
I’ve highlighted a few ways in which fishing maps can be improved. Of course, even with the best map, there is no guarantee that the big fish will take the lure.

Acknowledgements

This study has been conducted using E.U. Copernicus Marine Service Information; Global Ocean Physics Analysis and Forecast https://doi.org/10.48670/moi-00016 and Ocean Colour https://doi.org/10.48670/moi-00278. The study also used GHRSST-MUR data; JPL MUR MEaSUREs Project. 2015. GHRSST Level 4 MUR Global Foundation Sea Surface Temperature Analysis. Ver. 4.1. PO.DAAC, CA, USA. Dataset accessed [2023–04–29] at https://doi.org/10.5067/GHGMR-4FJ04. SST data were optained using NOAA ERDDAP; Simons, R.A., and Chris John. 2022. ERDDAP. https://coastwatch.pfeg.noaa.gov/erddap. Monterey, CA: NOAA/NMFS/SWFSC/ERD.