Mumbai: The Geo Artificial Intelligence and Random Forests based tools and technologies will predict and monitor air quality for Mumbai, Delhi, Bengaluru, Chennai, Kolkata, and other cities of India to help achieve Sustainable Development Goals (SDGs), informed Prof. P G Diwakar, ISRO Chair Professor, National Institute of Advanced Studies while addressing the gathering at India Clean Air Summit (ICAS) 2023, organised by the Centre for Air Pollution Studies (CAPS) at the Center for Study of Science, Technology, and Policy (CSTEP).
Fusion of Geo AI and Random Forests
Prof. Diwakar said that these computer-based tools are being harnessed in India to pool data from different sources. “We must link up SDGs and air pollution indicators, as highlighted by the United Nations 2030 Agenda. Particularly in the context of SDGs 3.9 that reflects on the health related issues and SDG 11.6 that reflects on the Cities and liveability indices in them. Using geospatial technology, we need to create well-distributed ground observation points to measure in-situ pollution parameters, which can further be integrated with space data for modelling. These multi-parametric sensors should be strategically placed to collect data from all representative points in a city or township.”
Mathematical Modelling for Comprehensive Estimates
He added that when we do mathematical modelling for seamlessly estimating pollution, on how air pollution can be addressed, it is necessary to have a geospatial outlook and modelling based on space and ground observations for Particulate Matter (PM), Nitrogen Oxides (NOx), (SOx), etc., and it could be linked to Aerosol Optical Depth (AOD), derived from Satellites sources, like, INSAT-3D/ 3DR, MODIs etc. “We need to bring in the weather data into the modelling framework, such as, wind speed, wind direction, surface air pressure, temperature, humidity etc, that play a major role as part of the model. The estimation of road fractions based on road type and carrying capacity in an urban setup is critical as they are the cause of a lot of pollution in urban setup, and all these could be well-integrated into numerical models to build forecasting and assessment. These have been tested for Bengaluru and very good results have been seen. We are also building Geo AI models that use classic Machine Learning and random forest theory, among others to provide not only current estimates seamlessly, but also can be used for forecasts effectively. These models that use data from ground and space can be very well integrated for a better life tomorrow,” Prof. Diwakar informed.
What is Artificial Intelligence?
AI is the ability of a computer to do tasks that are usually done by humans, and for this, they require human intelligence and discernment. Whereas, Random Forest is a machine learning algorithm, which combines the output of multiple decision trees to reach a single result. In both these cases, historical data becomes very essential. Already historical data of about 50 years has been compiled, basically from various satellite sources and we are also stacking all the possible ground based observations.
Pilot Project Initiated In Bengaluru
Detailing the project, Prof. Diwakar said: “We have initiated a pilot project in Bangalore and are actively assessing its effectiveness. Our aim is to extend this technology to predict and monitor air quality in heavily polluted cities such as Delhi, Mumbai, Chennai, Kolkata, and others. With unwavering confidence in our model’s capabilities, we will leverage historical data to predict and validate present air quality — an unprecedented feat in our country’s history. Our approach utilises conventional Artificial Neural Networks techniques, integrating diverse inputs from both space and ground sources to construct a robust predictive model. Spearheaded by the National Institute of Advanced Studies (NIAS), this project is set to be a collaborative effort among various groups.
Role of Geospatial Data
Emphasising the crucial role of geospatial data, the model will operate within a comprehensive geospatial framework. All data will seamlessly integrate into the meticulously curated geospatial database. “As progress is made, the focus will extend beyond air pollution to encompass broader SDGs, addressing issues such as water pollution, electromagnetic radiation, and more. This holistic initiative reflects the commitment to pioneering transformative solutions that transcend conventional boundaries,” said Prof. Diwakar.
Unifying for Cleaner Air
“Shedding light on India’s intricate airshed challenge, which possesses a diverse and complex structure, underscores the urgency of incorporating clean air objectives within our SDG framework (3, 9, 11, 17). The integration of clean air considerations is paramount, ensuring its prominence on our collective agenda. The essence of this summit lies in fostering a dynamic alliance among academia, industry, policy-makers, and civil society stakeholders. By recognising air pollution as a shared predicament, we strive for a unified solution. The current ecological landscape underscores the vitality of air quality. Framed within this context, our strategic approach must include precision in air quality assessment, dynamic modelling of environmental shifts, effective policy engagement that acknowledges air quality's impact, and the emerging focus on capacity building, nurturing expertise within our institutions and society,” said Dr. Pratima Singh, Senior Research Scientist, Air Quality, CSTEP.
