This objective of the Twenty-five Year Strategic Plan was largely a statement in support of the existing integrated catchment management initiatives established by the Queensland Department of Primary Industry in the interest of the catchment users themselves. The Plan spelled out the health of the GBR, the ultimate downstream sink, as yet another good reason for development of best practice in catchment use. Statistics do suggest there is significant potential for harm resulting from the cumulative effects of increased runoff of nutrients and silts into the Great Barrier Reef over the last century (Figure 4)¹⁴. However, as was pointed out by a QDPI land use manager in a rigorous round-table discussion with researchers: "If you want us to change the way people farm, you had better give us some ammunition", ie. The onus of proof was placed on the researcher and GBR managers.

Proof has so far eluded scientists. According to some researchers, phytoplankton blooms and seaweed abundance on some coral reefs are clear evidence of adverse effects caused by human populations on the coast and in the catchments¹⁵, though this is widely debated.

Some pertinent points are:

• Long-term studies by the Australian Institute of Marine Science (AIMS)¹⁶ have shown that 'green water' indicative of phytoplankton blooms, is a common natural occurrence in GBR waters. However, except within local and enclosed waters, it has not been possible to demonstrate that any human modification to runoff has significantly increased their intensity, frequency or duration, compared to the contributions of the natural recycling nutrient pool, plus nutrient imports from and exports to the north, east and south of the GBR.
• There are important regional differences along the 2,000 km length of the GBR. For example, the water is deep and the coral reefs are far from the Burdekin River, which has the biggest floods but many dry years. By contrast, reefs grow at the mouth of some of the more predictable smaller rivers north of Port Douglas.
• When a flood plume reaches a coral reef, the freshwater alone can kill the corals, molluscs and other animals, irrespective of the silts or nutrients it carries.
• The dead coral skeletons tend to be overgrown by seaweeds (plants which do not have roots and can attach to dead coral). But at reefs or parts of reefs where there are abundant fish or sea-urchins, the latter can eat the seaweeds almost as quickly as they grow, leaving the area bare except for sparse and wispy shoots of new growth¹⁷.

Where these grazers of the reef are absent or uncommon, seaweed growth can outrun the capacity of the grazers to eat it, and dense seaweed growth (anything from a centimetre to more than a metre high) can overwhelm the area. (In Jamaica, there is compelling evidence that overfishing of grazing fish and/or their predators can change the whole coral/seaweed balance on coralreefs¹⁸, but with its perceived relatively light fishing pressure, this is not considered a factor in Australia.) Other observations suggest that 'too many' fish or sea urchins can eat into the fallen rubble and/or reef structure itself (like soil erosion in an overstocked paddock) and that 'too much' phosphorus and nitrogen flowing into GBR waters can aid the seaweed growth and weaken the coral skeletons.

Moreover, while the seaweed can visually dominate an area for many years, it appears to represent one of two states:

• Indefinite occupancy where seaweeds exclude coral from settling or growing indefinitely;
• Transitional occupancy where a coral 'understorey' can develop, first noticed as saucer sized encrustations, but over a decade, growing into crowded tables and bushes that exclude the algae completely.

Seagrasses have roots and grow on sand and mud. Unlike seaweeds, they are considered 'good' because some of their type provide nurseries for fishes and prawns and/or food for the endangered dugong. However, they are also vulnerable to the effects of runoff. Interpreting changes in seagrass areas poses similar difficulties to interpreting increases in seaweed abundance. In 1992, over 1,000 km² of seagrass died in Hervey Bay due to storm waves (shallow seagrasses) and persistent light deprivation (greater than 10 m deep) caused by the turbid plume of a 100-year flood and resuspension of sediments in cyclonic seas¹⁹. This event led to starvation, deaths and emigration in an important dugong population that will take 25 years to recover²⁰. While 'runoff' was clearly a major cause of this event, it remains to be established whether agricultural or other land use in the catchments emptying into Hervey Bay caused the plume to be more turbid, to cover a larger area, or to persist longer than it would have, prior to the area's development for agriculture.

The seaweed and the seagrass stories recounted here thus do provide ammunition for the DPI catchment manager. The mechanisms by which runoff can do damage are understood, but we have been forced to acknowledge that 'other factors' (for example, existing nutrient pools, export and import of GBR nutrients, disturbance, and grazers as co-determinants of algal abundance) may equally influence ecological outcomes and management responses. The need to separate natural from people-generated changes remains unresolved.

Part of science's responses to this need is the continuation of scientifically unfashionable but important broad-scale ecological surveys and long-term monitoring programmes including the use of photo and video transacts in spatial-temporal monitoring (Figure 5)²¹. These show us 'how, and how fast' things change in different parts of the GBR, and stimulate ideas and research about the 'why'. The way forward for conveying geographically explicit information and advice to managers, currently under development, is seen to be through the use of advanced databases, knowledge-bases, Geographic Information Systems, risk assessment approaches and links to decision support processes.