 So we're getting better at being able to predict the weather. It saves money and it saves lives, but we're way behind when it comes to predicting other parts of the natural environment. But we need environmental prediction systems to help us predict biodiversity loss, prevent ecosystem collapse and to prevent even the entire collapse of human civilisations. Environmental prediction systems are critical given the pressure on the Earth's ecosystems and the risk of collapse is very important as happened on Easter Island about 400 years ago. But it's also important for assessing risks to business, assets and investments, particularly those that are associated with ecosystem function like oil palm plantations. But almost everything that we know about landscape change comes from pristine environments. But what happens when already altered environments are further altered? We know almost nothing about this form of secondary transformation. And this is critical given that more than 75% of the planet's surface is already altered. In the early 1990s I was travelling almost monthly to study the forests of Victoria. And as I was flying, below me thousands of hectares of grazing land was being altered and converted to pine plantations. He was an opportunity to study secondary transformation and primary transformation and help develop prediction systems. So we set up a 50,000 hectare experiment to quantify the impacts of landscape change on biodiversity and on the ecosystem itself and that experiment still running 23 years later where we've looked both at secondary and at primary transformation effects in those systems. And what we discovered was that secondary transformation leads to novel combinations of species. We also discovered that secondary transformation has markedly different effects to primary transformation. But most importantly and very exciting was that it was possible to predict the effects of those changes and with a good degree of accuracy. But those outcomes are for one ecosystem and one set of changes. What about other ecosystems? Can we predict the changes in them too? We have other large scale long term studies with different drivers of ecosystem change. Fire, logging, agriculture and restoration. And the answer was yes, we were able to predict the changes and those effects. Importantly there was some general lessons from those big studies. The first one was is that we needed to have long term data sets to develop our environmental prediction systems. We can't predict in the future if we don't know where we've been in the past. The second thing was that environmental prediction systems need to be based on high quality data sets that focus on key parts of ecosystems. And some of those parts of ecosystems need to be retained when landscapes are undergoing transformation. Habitat patches, large old trees, rocky outcrops are some of those examples. And often they're quite small and sometimes overlooked but they're critical what we call the Frodo effect from Lord of the Rings. Long term monitoring data are critical to develop environmental prediction systems but monitoring is regarded as a Cinderella science, underpaid, undervalued, the first thing cut, the last thing funded. And we argue that we need to change this if we're going to make progress. But there are huge opportunities for gathering the data that the world needs particularly through large scale environmental programs like those restoring millions of hectares of degraded land. The problem is that those opportunities are wasted and with it billions of dollars of environmental spending every year. Another way to secure monitoring systems is through environmental accounting. That allows us to make better decisions and to assess risks to economic assets and changes associated with environmental change. And that's exactly what we've done in these magnificent forests of Victoria, the tallest flowering plants in the world. And accounting has showed us how to assess the economic value of different natural assets, for example native timber versus tourism versus water supply. It's also allowed us to quantify the trade-offs associated with environmental decisions and management change. And this allows us to make better business decisions and better environmental policy. The world needs environmental prediction systems more than ever given the pressure on ecosystems and the risks of ecosystem collapse. But we need major data sets through long-term environmental monitoring to be able to develop those prediction systems. So there are huge opportunities to build monitoring into environmental programs. In fact, I think all long-term environmental programs must be properly monitored and we can add value to those monitoring data by putting them into formal environmental accounting systems. And I think we have to take these opportunities if we're going to tackle the problems with environmental change on the planet. Thanks very much.