Data-Driven Environmental Solutions
The impact of Data Science and AI on the future of Environmental activities in Iran
Interview with Professor Soroosh Sorooshian By Ehsan Daneshvar
Ehsan Daaneshvar: Why is ensuring data and datasets' accuracy and reliability in environmental science critical?
Soroosh Sorooshian: On the importance of the accuracy and reliability of data and datasets in environmental science, it is of paramount importance that the quality of the data observations is continuously checked for accuracy and completeness. It is the quality of the data that will determine the usefulness of the information for decision-making. No realistic and effective decision about managing our environment (water, air, eco-system, etc.) can be made without information about the physical behavior, historical trends, and patterns of these Earth system elements. Useful information for planning and decision-making, which could be the by-products of model simulations, and statistical analysis of trends and patterns, can only be generated from raw observations. In the case of our climate and water resources, we rely on the historical time series of our environment, which are dynamic and evolve over time, such as river flows, precipitation (rainfall and snow), and air temperature, among others.
E.D.: What are the strengths and weaknesses of available datasets, current systems, and how modeling is conducted in Iran concerning environmental sciences and analysis?
S.S.: Perhaps I am not the best person to reflect on this question since I am not directly involved with any project in Iran, which provides me with insight into the data collection systems and how well they are maintained and quality controlled. However, judging from numerous publications by our Iranian colleagues related to the hydro-climatological studies of Iran, the data seems to be of reasonable quality. The only issue I have observed has been some gaps in the data time series of rainfall over Iran, especially during the period of the Iran-Iraq War, which disrupted programs related to environmental monitoring.
It is worth noting that state-of-the-art instrumentation has advanced in the past several decades. Among them are ground-based meteorological radars to observe high-resolution rainfall, radar-based water level sensors to measure river flows, as well as other commercially available monitoring systems. It is prudent that any nation keeps up with environmental monitoring to the best possible extent possible.
Regarding the modeling component of the question, again, I am not as familiar as I should be. However, the Iranian engineers and scientists working in the environmental field are generally well-educated and familiar with many of the state-of-the-art hydrologic, hydraulic, and hydrometeorological models available to the global scientific community without any restrictions.
E.D.: Tell us about data and environmental monitoring in EMP of water resources structures development plans.
S.S.: For most parts, environmental management plans (EMPs) are specific to the system being managed and monitored. In the case of air pollution and air quality, international standards have been established used to monitor the health and well-being of residents of urban areas. Fortunately, standardized monitoring instruments are widely available and collect the data needed for implementing EMPs.
In the case of water resource variables crucial for management, EMPs will likely require customization to the condition of each case. For instance, standards applied to one water resource structure(ex. a reservoir) will need to be conditioned for each specific reservoir, depending on their operating rules and regulatory requirements. Therefore, the availability of observed data and continuity will be critical to developing EMPs and operating standards.
E.D.: What revolutionary and radical measures must be undertaken in the future to enhance the capabilities of model development through reliable data acquisition, processing, and interpretation?
S.S.: What is important is to follow perhaps the example of countries such as China, as well as many other nations, and establish modeling centers with the ability to develop and maintain environmental models useful for regional planning. International collaboration will go far in creating a win-win situation in developing regionally customized approaches that will benefit the region and provide the required information and linkage to global modeling and observational activities.
E.D.: How big data and ML algorithms would help environmental studies to reduce emissions and develop master plans?
S.S.: Undoubtedly, the advances in machine learning (ML) and artificial intelligence (AI) have revolutionized the ability to provide useful information for modeling and predicting many complex systems, such as our environment. However, success depends on the availability of big data needed for these algorithms to succeed and establish patterns and system behaviors. Without adequate and reliable environmental observations and long-term series, one cannot expect such tools to succeed. Unfortunately, with the ever-increasing number of profit-driven scientific journals, there has been a proliferation in recent decades of research articles in this area. The more unfortunate matter is the quality of the review process has diminished significantly, and many of these articles are nothing more than click-bait with questionable content. I have seen papers that claim to have discovered doomsday scenarios of extreme climatic and ecological shifts while using inadequate data to substantiate their claims. In other words, the field of environmental sciences is not similar to the Amazons of the world simply because our “big-data” requirements are not similar to what drives their success in following customer behavior, etc.
E.D: What are the roles of AI and data science in the future of Iran’s environmental studies/ development/ sustainability?
S.S.: This is an excellent question. There is no doubt that AI and data can be extremely beneficial in providing information for environmental applications when used properly and responsibly. However, one must appreciate that these tools are limited to providing useful information. By analogy, it’s inefficient to have a semi-truck be used as a taxi to carry passengers around. These methods are sophisticated and can provide meaningful information as long as their required data has been collected and processed. In several areas of our environmental systems, remote sensing has revolutionized the field of earth systems, as we can now access high-resolution observations that can be used with AI and data science algorithms to produce valuable information such as rainfall. Our center here at UC Irvine uses neural network algorithms to convert remotely sensed observations to precipitation estimates over the entire world at high resolution in both real-time and for long-term hydroclimate studies of extremes and trends. We have seen exponential growth in the download and use of our data products by Iranian scientists and engineers conducting studies and developing their own unique applications. In fact, Iran ranks in 7th place among over 200 nations and territories in the use of our data.
E.D.: How can we mitigate the environmental impacts of future development plans in general?
S.S.: It requires social scientists and economists to address the complex issues at the heart of this question. Of course, the availability of information about environmental patterns and trends is required, which can be supplied by the scientific and engineering communities studying and monitoring the environmental systems.
Professor Soroosh Sorooshian is the Director of the Center for Hydrometeorology & Remote Sensing (CHRS) and Distinguished Professor of Civil & Environmental Engineering and Earth System Science Departments at UC Irvine. Prior to 2003 he was a faculty at the University of Arizona for 20 years. His area of expertise is Hydrometeorology, water resources systems, climate studies and application of remote sensing to earth science problems with special focus on the hydrologic cycle and water resources issues of arid and semi-arid zones. He also consults on problems related to surface hydrology and urban flooding.