West Sydney Curbs Heat with Smart Irrigation

The Sydney Olympic Games were, for many, the best games ever. The sports event left a legacy of inclusion, urban development, a new international position for the city and above all, sustainability and environmental responsibility. For the games, the city developed a rundown area in the West, which has now attracted many residents, and created the Olympic Park on top of wasteland, giving it a second life. But because it was built over a very thin layer of soil, the park dries out very quickly, meaning the irrigation system is crucial to keep the area healthy. One of the parks inside this green complex is the Bicentennial Park, which in 2019 was deeply impacted when nobody realized that the power had gone down and with it, the irrigation system. That's when the Bicentennial Park decided to turn to smart irrigation. But with smart irrigation came the opportunity of doing a duathlon: irrigation and curbing urban heat island effects. We're now accelerating this heat problem. You have a lot of people living in an area that gets really, really hot naturally. And now, of course, the urban development that we see means that we are capping land. We're transforming vegetated open surface to hard surfaces. And our idea was to look at means how we can reverse what then leads to urban overheating and the urban heat island effect. And in the scientific literature, what is really well established from many places around the world is the park cool island effect. West Sydney has a population of around 2.5 million, a number expected to grow by 1.5 million in the next ten years. So urbanization is in the cards. The park saw having a smart irrigation system as an opportunity to make better watering choices and lower temperatures, which is why SIMPaCT was created. 12 partners, including universities, private companies, governments, organizations and even the utility Sydney Water, came together to develop and deploy a smart irrigation system to keep spaces green and cool at the Bicentennial Park, a 40 hectare space with 200 irrigation zones. We used about 200 soil moisture sensors to make sure that we have a single sensor in each of those zones. We also had a total of 13 complete weather stations that recorded not just temperature and relative humidity, but also the amount of precipitation, wind speed, wind direction, lightning strikes, and so on and so forth. And we put seven of those stations in the park and six of those stations about a kilometer away into the urban core. So we could compare. And we do compare. The data that comes from the park, how warm or cold is it there, against the urban precinct. How warm or cold is it where the people are? The 200 soil moisture sensors and 50 temperature and humidity weather sensors are by company Sensedge and the 13 weather stations from Lufft. The soil sensors use Lorawan and connect to three gateways, two in the park and one on a nearby skyscraper. And the stations use 4G cellular connectivity. The data, gathered every 15 to 60 minutes, then runs into one of the stars of the project: the digital twin of the park. So for us, it's a fully automated process. So basically we are streaming data coming in in real time, but we are then usually running forecasts. What's happening for a given day at six in the morning. So it's kind of this is one when we're looking into what is the irrigation that has been scheduled and executed overnight. What are the current soil moisture conditions. And then we’re also getting in updated meteorological data. So for us to Bureau of Meteorology, in Australia is producing those data. And then we take that all in and then it's automatically looking into the thousands of different possible irrigation scenarios. So this is where your machine learning and optimization algorithms come in. So this is also fully automated. And based on that information then the digital twin is making a decision how much water should be scheduled on that given day to be irrigated overnight. Considering the forecast, nowcast and hindcast, which are analyzed in the digital twin through machine learning algorithms, the irrigation recommendations go through Senaps, which sends the schedule to the irrigation program dubbed FieldMouse, in charge of the irrigation software Irrinet, responsible for opening and closing the water valves. The digital twin also analyzes the use of water and compares it to the data from the moisture sensors, calculating if the amount is correct and spotting leaks if it's not, meaning the solution is also bringing an efficient use of water. We were estimating a reduction in water use of about 15 to 20%. Once we have the algorithm really learning and understanding what it's doing, it can just use as much water as we need, and we don't need to be on the safe side anymore because the machine learning algorithms have learned, the AI has learned what it actually takes to keep a certain area hydrated in a way that the vegetation is just happy during winter and during the summer, we get the best cooling off of it. I have measured that our park environment can be 7 degrees cooler than the urban environment, and that, of course, means the cooler air that is produced in that park through irrigation and therefore optimal transpiration from the plants is then transported into the surrounding environment. The data is made available to the public through a web page for them to see what the temperatures are like in the park, and then be able to decide where to go. The project started in 2021 and had a budget of around $6 million. It can be seen as a high budget, considering only $350,000 went to hardware, the rest went to talent, software development and all the specialized different actors involved in the project, due to all the knowledge needed, among others. However, we've seen these high budgets in other projects that are new and research, and in which talent takes a big chunk of the budget, like FloodNet in New York City, which had a cost of about $7 million with 500 sensors. For SIMPaCT, $2.5 million came from the Digital Restart Fund, 3 million from the Sydney Olympic Parks, universities and private companies, and half a million came from the utility Sydney Water, who will see the return on investment by learning how they can increase water usage efficiency for cooling and then replicate their solution. Because everything is internet based, we could irrigate parks in Tokyo or in Valencia or in Sao Paulo. It doesn't really matter for us. All we need to do is develop the adapter so that the language that SIMPaCT speaks, the digital language for those irrigation controls can be translated into irrigation controls that the irrigation system in whatever city or whatever place speaks. We are talking to people in Dubai, we are talking to people in Europe. We're also talking to people, of course, here in Sydney. What makes this project different from others, though, is that it tackles water and heat simultaneously through the use of machine learning and a digital twin to generate predictions and schedules not only with IoT data, but data from other sources, which would take a long time to analyze without AI. Tackling both pains can be a game changer. On the one hand, we have the idea of having more green spaces to curb heat. On the other, we have water wastage, which can be a worry when incorporating more green areas. However, while water continues to be a concern, data solutions for irrigation and heat can encourage more cities to create greener spaces as they can be assured that no water will be wasted while cooling the environment and keeping the plants lush.