Are farmers to blame?

The Agricultural hypothesis

A short account of the life cycle of CoTs provides the grounds for looking in detail at the proposed link between agricultural fertilizer and CoTS plagues. CoTS live for around four years. After about two years they can reproduce tens of millions of microscopic larvae that float around in the water being carried in whatever the direction of ocean currents. After about two weeks to a month, the larvae settle on a suitable surface on the seabed and metamorphose eventually to become a small starfish. In this time, they could have travelled many hundreds of kilometres in the ocean current but, once they settle, their wandering days are effectively over.

At first, they do not eat coral as they are far too small. Instead they feed on organisms such as the hard coralline algae that makes up a substantial fraction of any coral reef. As their size increases, they become coral-eating machines favouring the fast growing “plate” and “staghorn” type corals. They do not physically bite into the coral. Instead they digest the polyp at the surface of the coral as they move over its surface. The dead coral skeleton remains.

Despite half a century of research it is surprising how little is known about these starfish. As already mentioned in the major review paper published in 2017 (Pratchett et al, 2017)[1], there is considerable debate about whether CoTS were in plague proportions before the 1960s. But there are many other important scientific gaps. For example, very little is known about the mortality rates, or even what causes mortality, of the larvae and juveniles. This is important because the agricultural hypothesis rests on the proposition that mortality is reduced by the increase in nutrients from fertilizer. With such little idea about mortality rates of CoTS, it is unwise to be too quick to blame the farmers.

The 2017 Scientific Consensus Statement claimed, concerning the link between agriculture fertilizer and CoTS, that

  1. The CoTS outbreaks always occur around three years after a major river flood. These outbreaks start around Lizard Island, well to the north of the major river mouth (Burdekin River) in one of the most pristine and remote parts of the Reef and hundreds of kilometres from the nearest area of significant agriculture  (see figure 3.3).
  2. Laboratory experiments demonstrate that larval survival increases dramatically with a small increase in phytoplankton concentration. At the very low concentrations found in far-offshore water survival of CoTS larvae is almost zero, but in concentrations found closer to shore, which is claimed to be elevated by agricultural fertilizer, survival is between ten to a hundred times higher (Fabricius et al., 2010).[2]
  3. Observations of phytoplankton concentration in the Central Section of the GBR, which is adjacent to farming regions, is twice as high as for the areas in the Northern Section of the GBR where there is almost no agriculture (Brodie at al., 2007).[3] This claim is used to demonstrate that there is a long-term increase in the phytoplankton caused by agriculture.

Each of these three claims struggle when scrutinised.

Figure 3.3 Map of the GBR showing Lizard Island where it is claimed that CoTS plagues begin. Nutrients from river floods are claimed to be the reason for the plagues. The most important river , the Burdekin, is around 600 km from Lizard Island. The Swains Reefs are also marked and are the area of the most persistent plague of CoTs, but are furthest from the coast and agricultural influence.

Link between river floods and CoTS outbreaks

Figure 3.4 Estimated annual river discharge (total bars) from 1949-2012 wet seasons for north and central GBR. Dark bars are early wet season flows (Nov-Feb). White bars are late wet season flows. Arrows indicate nominal years of CoTS outbreak initiation. Data from Devlin et al. (2013).[5]

Figure 3.4 shows the typical argument used to demonstrate the link between river floods and CoTS outbreaks.[4] The dates of the four major outbreaks at Lizard Island are marked with arrows, three of which occur roughly three years after years of major floods. As it takes three years for CoTS to reach maturity, we must expect a three-year lag between the flood and the initial outbreak. Because the CoTS spawn sometime after November, which is at the end of the dry season, the data are broken into early and late wet season with the early wet season data being more significant because the CoTS larvae are only viable for a few weeks. A flood many months after spawning could have no impact as the larvae would already have settled and become juvenile starfish.

The relationship in Figure 3.4 looks reasonably convincing especially considering that environmental data rarely produce beautiful correlations such as those in physics and chemistry under carefully controlled laboratory conditions. The actual world tends to be more complicated and messier so it is unlikely that data are a perfect match. Even so, there are many problems with this implied relationship.

The first problem is that only three of the four known major outbreaks occur after major floods. The outbreak in the 1960s does not fit. The second is that the lag between the flood and the outbreak is not consistent, being around five years for the 1979 outbreak, which is too long, and between one and three years for the latter two outbreaks.

It is perhaps a more important problem that the river data includes the Burdekin River. Because the Burdekin has a greater discharge than all other rivers in the region,[6] the Burdekin data dominate the river discharge volumes. For example, the Burdekin discharge for the record-year 1974 was twice as much as all the other rivers in the region combined. This is a problem because the Burdekin River is around 600 km from Lizard Island where the outbreaks commence (Figure 4.3). So the question must be asked – why do the outbreaks not start much closer to the Burdekin? The nutrients from the Burdekin would be in far higher concentrations, say, 200 km to the north where the Reef comes close to the coast. But the outbreaks only seem to occur at an extreme distance. (To use an analogy in air pollution, it is equivalent to arguing that there is more effect on the environment from air pollution very far from the source than close by.)

A third problem is that there are only four data points (only four recorded plagues of starfish) with which to make some sort of correlation or relationship, but a large number of variables to consider about the river discharge, river distances, flood timings. The variables include not only the combined flow of all the rivers in North Queensland, but also the individual rivers, and various combinations of the various rivers. In addition, the scientists could vary the boundary date between early wet season and late wet season plus the threshold value of annual flood volume which in this case was set to a discharge of 10 km3.

 So, at the minimum, when the scientists were looking for a relationship, they had four parameters at their disposal to produce a relationship, and four data points to test the relationship. With enough effort and torturing of the data, with only four events and (a) highly variable flood data, (b) the ability to select a wide range of rivers sometimes at extreme distances, and (c) a wide range of choices for selecting the boundary between the early and late wet season and the threshold flood volume, it is almost inevitable that some sort of loose correlation between CoTS and river discharge could be built.  As it is, the relationship only fits three events.

A fundamental principle in mathematics or modelling is that there should be far more data than variables in order to attempt to find a statistical relationship.[7] The brilliant mathematician and physicist, John von Neumann, was reputed to have said to the equally brilliant physicist Enrico Fermi

with four parameters I can fit an elephant, and with five I can make him wiggle his trunk.[8]

Von Neuman was pointing out that with enough variables, and without much data, one can produce a “model” that can be varied to give almost any result that is required. The CoTS hypothesis undoubtedly contains an infamous “von Neumann Elephant.”

There are other observations that do not fit the link between CoTS and rivers. For example, there was a continuous 20–year long outbreak in CoTS between 1985 and 2004 in the Swains Reefs[9] at the southern end of the GBR and most distant of all reefs from the mainland and completely unaffected by river discharge. No other part of the Reef has been recorded to have such persistently high number of CoTS for such a long period.

This means that river discharge cannot be the whole story.

It is also peculiar that, if CoTS rely upon river floods, why do they spawn in the period that spans the late dry season into the early wet season, that is, November to  February. November is one of the driest months of the year and the wet season does not usually start until the end of December.  In order to align with floods, the best spawning period would be properly aligned with the wet season from January to March.

It is notable that CoTS plagues have been documented elsewhere in Australia and the world where the river influence is negligible, casting further doubt on the link between agriculture and CoTS plagues. For example, plagues have been documented on coral reefs adjacent to desert areas of Western Australia where river discharge is minimal and agriculture is almost non-existent.[10]

It should thus be evident that although there is a potentially interesting link between rivers and CoTS outbreaks, it remains an interesting hypothesis to be explored rather than a well proven fact.

The review by CoTS experts (Pratchett et al., 2017) stated that the question of whether

high nutrient conditions and associated phytoplankton bloom  . . . coincide with observed spawning period . . . was Largely Unresolved.

It is thus surprising that the 2017 Scientific Consensus Statement concludes the opposite and that farmers are to blame.

Does CoTS larval survival increase with higher phytoplankton concentrations

The 2017 Scientific Consensus Statement relied upon the laboratory work of Fabricius et al. (2010) to draw the link between high phytoplankton concentrations and CoTS larval survival. Fabricius et al. (2010) found that increasing the phytoplankton concentrations by a factor of two could increase the larval survival by a factor of 10 or more. Other authors have found completely different results, however. For example, Wolfe et al. (2017)[11] found only a small change in larval survival across a wide range of realistic phytoplankton concentrations. Quite remarkably this contradictory work of Wolfe et al (2017) was cited in the 2017 Science Consensus Statement to agree with Fabricius et al. (2010), that is, it implied that the findings of Wolfe et al. (2017) were completely the opposite of what was actually stated by Wolfe. This is one of a very large number of difficulties with the 2017 Scientific Consensus Statement.

The fact that CoTS outbreaks occur in regions all over the world where nutrients and phytoplankton concentrations are very low is also suggestive that these factors are not essential for plagues and that other factors must be important.

Answering the question of whether high nutrient conditions needed for the enhanced survival of larvae in the field, the CoTS experts, Pratchett et al. (2017) concluded that

while receiving considerable attention, these questions are Largely Unresolved.

The Scientific Consensus Statement nevertheless blames the farmers.

Do regions adjacent to agriculture have consistently higher phytoplankton concentrations?

There is not much doubt that during periods of river flooding, the nutrient concentrations in the river plumes are higher than it would have been before European settlement due to agriculture (for more detail, see chapter 6). These river plumes last a few days per year and will have caused elevated phytoplankton concentrations for the period that the plumes existed. However, it is also often claimed that the influence of this short riverine input of nutrients lasts throughout the year causing consistently doubled concentrations of phytoplankton, especially close to the coast (Scientific Consensus Statement 2017 p18; Brodie et al., 2007; Fabricius et al., 2010). It has been claimed (Fabricius et al., 2010) that this contributes to the growth in CoTS numbers over long periods of time.

This issue is addressed more fully below (chapter 7), but the claim that phytoplankton concentrations have doubled is very doubtful due to systematic biases in the measurement program that was used to justify the claim (Brodie et al., 2007). There is also a very good oceanographic reason why consistently high phytoplankton from riverine nutrients is virtually impossible. This is because enormous quantities of water with low nutrient levels flow into the GBR from the Pacific Ocean. Finally, the dominant factor in the nutrient budget, aside from water exchange with the Pacific Ocean, is the cycling of nutrients across the seabed which are at least 100 times greater than that delivered from rivers.[12]


The link between agricultural nutrients and CoTS plagues as proposed by the 2017 Scientific Consensus Statement is tenuous in the extreme and confidence in the Consensus Statement is further reduced because it fails to mention the multiple problems with this hypothesis. These omissions include the geological record that demonstrates CoTS have occurred in large numbers for thousands of years, and the occurrence of CoTS plagues in low nutrient water in the GBR, and many other parts of the world, where there is no river influence at all.

The Scientific Consensus Statement also failed to mention that there is a significant body of research and other studies that challenges the claim that high nutrient and phytoplankton concentrations are required and, in one important case, incorrectly cite this contradictory evidence as supporting their hypothesis (Wolfe et al., 2017).

Stringent regulations have been placed upon the agricultural industries adjacent to the GBR including reductions in fertilizer application that will likely affect productivity and profitability. There is a very high chance that this is an expensive solution to a problem that does not exist.

[1] Pratchett, M.S., Caballes, C.F., Wilmes, J.C., Matthews, S., Mellin, C., Sweatman, H.P.A., Nadler, L.E., Brodie, J., Thompson, C.A., Hoey, J., Bos, A.R., Byrne, M., Messmer, V., Fortunato, S.A.V., Chen, C.C.M., Buck, A.C.E., Babcock, R.C. and Uthicke, S. (2017). Thirty Years of Research on Crown-of-Thorns Starfish (1986–2016): Scientific Advances and Emerging Opportunities. Diversity, [online] 9(4), p.41. Available at:

[2] Fabricius, K.E., Okaji, K. and De’ath, G. (2010). Three lines of evidence to link outbreaks of the crown-of-thorns seastar Acanthaster planci to the release of larval food limitation. Coral Reefs, 29(3), pp.593–605.

[3] Brodie, J., De’ath, G., Devlin, M., Furnas, M. and Wright, M. (2007). Spatial and temporal patterns of near-surface chlorophyll a in the Great Barrier Reef lagoon. Marine and Freshwater Research, 58(4), p.342.

[4] Devlin, M., Petus, C., da Silva, E., Alvarez-Romero, J., Zeh, D., Waterhouse, J. and Brodie, J. (2013). Chapter 5: Mapping of exposure to flood plumes, water types and exposure to pollutants (DIN, TSS) in the Great Barrier Reef: toward the production of operational risk maps for the world’s most iconic marine ecosystem. In: Assessment of the relative risk of water quality to ecosystems of the Great Barrier Reef: Supporting Studies. A report to the Department of the Environment and Heritage Protection, Queensland Government, Brisbane. TropWATER Report 13/30, Townsville, Australia.

[5] See Endnote 14.

[6] Furnas, M. (2003). Catchments and Corals: Terrestrial Runoff to the Great Barrier Reef. Australian Institute of Marine Science, Australian Institute of Marine Science, p.350.

[7] Dyson, F. (2004). A meeting with Enrico Fermi. Nature, 427(6972), pp.297–297.

[8] See Endnote 17

[9] Lawrey, E. (2013). Crown of Thorns Starfish (COTS) outbreaks on the Great Barrier Reef [animation] | eAtlas. [online] Available at:

[10] Haywood, M.D.E., Thomson, D.P., Babcock, R.C., Pillans, R.D., Keesing, J.K., Miller, M., Rochester, W.A., Donovan, A., Evans, R.D., Shedrawi, G. and Field, S.N. (2019). Crown-of-thorns starfish impede the recovery potential of coral reefs following bleaching. Marine Biology, 166(7).

[11] Wolfe, K., Graba-Landry, A., Dworjanyn, S.A. and Byrne, M. (2017). Superstars: Assessing nutrient thresholds for enhanced larval success of Acanthaster planci , a review of the evidence. Marine Pollution Bulletin, 116(1–2), pp.307–314.

[12] Furnas, M., Alongi, D., McKinnon, D., Trott, L. and Skuza, M. (2011). Regional-scale nitrogen and phosphorus budgets for the northern (14°S) and central (17°S) Great Barrier Reef shelf ecosystem. Continental Shelf Research, 31(19–20), pp.1967–1990.

%d bloggers like this: