Combining Advanced Sensors and Big Data with Artificial Intelligence to Understand the Response of Lake Ecosystems to Climate Change

Project Title: Combining Advanced Sensors and Big Data with Artificial Intelligence to Understand the Response of Lake Ecosystems to Climate Change

Project Lead’s Name: Craig Williamson


Phone: (513) 529-3180

Please Choose the Primary Affiliation: CAS

Are There Other Project Team Members?: Yes

Other Project Team Member: Allison Jones-Farmer

Other Team Member Email Address:

Other Project Team Member: Jing Zhang

Other Team Member Email Address:

Brief description of project: The Miami Big Data Initiative (MiEBDI, is a new initiative in the College of Arts and Science that facilitates student and faculty access to, and analysis of, ecological big data from around the world to address critical threats to ecosystems and the valuable services that they provide. Miami University's Center for Analytics and Data Science (CADS, is a university-wide initiative bringing together leadership from Miami's Farmer School of Business, College of Arts and Science, College of Engineering and Computing, and others to foster collaborations between students, faculty, and staff working with analytics and data science. Both groups have common goals to prepare Miami undergraduates and graduate students to work with big data, and to foster interdisciplinary research among students and faculty to provide opportunities for publication and external funding centered on the analysis of big data. Here we request funds for a suite of advanced sensors that will serve as a platform for a collaboration between students and faculty in CADS and MiEBDI through the creation of a big data learning module. We will combine two emerging frontiers: (1) the use of advanced environmental sensors and the big data that they generate in MiEBDI with (2) the use of artificial intelligence and advanced analytics to give students experience applying these tools to address grand challenges in global ecology related to managing freshwater resources. The cutting edge question that we will address is how climate-change induces oxygen depletion in lakes, a recent hot topic that has been addressed in oceans, but not in lakes (see February 2017 Washington Post article and Nature paper: Undergraduates from multiple disciplines will be involved in this project, including those from Biology, Statistics, and Information Systems & Analytics. The proposed collaboration will provide students with the skills necessary to work with ecological big data (or any big data for that matter) and sensors, and experience using artificial intelligence to understand these complex data sets. Members from several departments have expressed support for and interest in being involved with the development of this learning module, including Jing Zhang and Tom Fisher (Statistics, CADS), Allison Jones-Farmer, Maria Weese, and Arthur Carvalho (Information Systems and Analytics, CADS), Craig Williamson, Mike Vanni, and Beth Mette (Biology, MiEBDI). Courses that will be able to use the sensor data from this learning module and benefit from the advanced analytics include the following:

Course | Approx. Students per year
Environmental Protocols IES 411/511 | 15

Limnology BIO 463/563 | 24
Introduction to Hydrogeology GLG 408/508 | 25
Hydrogeography GEO 425 | 10
Hydrogeologic Modeling GLG 428/528 | 7
Field Hydrogeology GLG 499/599 | 6
Microbial Ecology MBI 475 | 25
Analysis of Forecasting Systems STA 483/583 | 24
Managing Big Data ISA 414 | 50
Modeling and Study Design STA 672 | 20
Total Students per year: 206

Many students conducting research projects will also benefit from using the sensor data and advanced analytical approaches. Real-time data gathered from these sensors will be used in a CADS
"sandbox" project starting in the Fall of 2017 which will be comprised of 6-8 undergraduate students from multiple disciplines. Faculty directed internship opportunities involving corporate outreach will also be a part of the proposed project, including IBM Corporation, and two Ohio-based companies: Fondriest Environmental, Inc., and NexSens, Inc. The two Ohio-based companies specialize in environmental sensors including the sensors that we are proposing to use in this study, and have expressed an interest in being involved in this collaboration between CADS and MiEBDI. In past collaborations with Fondriest and NexSens both undergraduate and graduate students have participated in internships to further their education through understanding sensor technology. The proposed project has the potential to provide Miami students with similar opportunities through all three corporations.

Does this project focus on graduate student education or graduate student life?: No

Describe the problem you are attempting to solve and your approach for solving that problem.: Our changing climate is having profound impacts on our planet. The effects of climate change are manifested, in part, by increases in extreme weather events. As the lowest point in the landscape lakes are highly responsive sentinels of climate change, responding to regional to global scale changes. Lakes exhibit physical, chemical, and biological responses (e.g. changes in water transparency, temperature, dissolved oxygen, dissolved organic matter, and chlorophyll) to climate change as well as shorter-term weather events. Lakes in northern regions are more sensitive to climate change due to shorter ice cover duration and longer ice-free periods during the summer. Two of the most pronounced responses of lakes to climate change are changes in thermal structure (water temperature at different depths) and oxygen. As with marine systems, human activity is creating "dead zones" where oxygen depletion occurs in deeper waters. Climate change is aggravating this depletion of oxygen in lakes and oceans, extending and creating new dead zones that threaten fish and other aquatic life. This creates an urgent threat to aquatic ecosystems and the potential to use advanced sensors in lakes as a learning module for our students to collect, analyze, and apply big data to help address this global challenge. We propose to deploy PME miniDOT temperature/ dissolved oxygen sensors (Fondriest Environmental, Inc.) throughout the water columns of Lakes Lacawac and Giles (6 sensors per lake, 12 sensors total) for 12 months of each year. These sensors will collect high-frequency data (every 15 minutes) on temperature and dissolved oxygen that will serve as a core platform for the learning module and collaboration between CADS and MiEBDI students, who will work to link these data sets with nearby weather data to assess the effects of weather events and climate change on lake ecosystems. These resulting big data sets will allow us to assess how the temperature and oxygen dynamics of the lakes respond to various weather conditions over the course of a year, capturing events that cannot be recognized through traditional manual (weekly to monthly) sampling. Application of artificial intelligence and advanced analytical approaches from CADS and long-term lake data available through MiEBDI will be used to develop predictive models of how lakes respond to climate change, leading to new insights. The primary products of the proposed collaboration will be more effective predictive models of water quality, and more importantly, the training of our students across disciplines for the future challenges in applications of big data science to managing freshwater resources. Summer as well as academic year internships and educational opportunities will be available to undergraduate students. One that we are currently exploring is with the IBM Corporation supported by an existing partnership through CADS. Through this partnership, students and faculty will leverage IBM cloud, cognitive computing resources and weather data to explore new research questions with a major focus on using this platform of lake temperature and oxygen data. We also have collaborations with two Ohio-based companies (Fondriest Inc., and NexSens Inc.) that manufacture and sell advanced sensor systems for aquatic ecosystems. Our undergraduates and graduate students have done internships at these companies in the past and these companies have expressed an interest in getting more actively involved in the collaboration proposed here with the oxygen-temperature sensors. We also have extensive data (20+ years) on the two Pennsylvania lakes in protected watersheds (good climate indicators), Lake Lacawac and Lake Giles, where the sensors will be deployed. The Lacawac Field Station provides housing and a new NSF-funded laboratory that are available for use by Miami students at no charge due to Miami's membership in their consortium. Lacawac also offers summer internships for our students. A subset of the CADS-MiEBDI interns will have the opportunity to engage in the hands-on deployment and data collection in the field, providing the data for the larger group of students in academic year internships and classes that will use the data.

The criteria state that technology fee projects should benefit students in innovative and/or significant ways. How would you describe the innovation and/or significance of your project?: This project is unique in that it will bring together the diverse expertise of members of CADS and MiEBDI as well as several diverse departments, Biology, Statistics, and Information Systems and
Analytics, to uniquely address the emerging frontiers of sensor technology and ecological big data using artificial intelligence. This range of expertise will both advance our understanding of the
effects of climate change on lake ecosystems, but more importantly train our students with innovative approaches to the collection and analysis of big data. This will benefit the students involved in the project by exposing them to approaches and expertise beyond that which they would typically receive in their normal undergraduate and graduate programs. The students will gain experience with ecological sensors and big data, and the unique opportunity to combine these emerging frontiers in ecology through MiEBDI with the innovative opportunities in artificial intelligence within CADS.

How will you assess the project?: The project will be assessed through several mechanisms. We will keep track of the number of students who do internships with both CADS and MiEBDI, and give them exit questionnaires about their experience including how much they learned about both the use of advanced sensors in environmental sciences and the use of artificial intelligence to improve predictive lake modeling. We will assess their success in being admitted to good graduate schools and/or finding high-level jobs. We will also keep track of the number of undergraduates that attend the MiEBDI data workshops that utilize these sensor data as examples, and keep track of the number of classes that employ the resulting data and findings in teaching about environmental science. The success of the researchers (students and faculty involved) at using cognitive computing to understand lake responses to weather events, and applying this information to models of lake dynamics, will also be assessed through co-authorship of undergraduate and graduate students as well as associated faculty on peer-reviewed publications, conference presentations, and presentations at Miami's Undergraduate Research Forum.

Have you applied for and/or received Tech Fee awards in past years?: No

What happens to the project in year two and beyond? Will there be any ongoing costs such as software or hardware maintenance, supplies, staffing, etc.? How will these be funded?: These sensors will last well beyond a year. They will be maintainedand serviced by the Center for Aquatic and Watershed Sciences of which Williamson is an active member. CAWS has one full-time employee, Research Associate Tera Ratliff, who has a Master's Degree in Biology and extensive experience with sensors and data management. She is a permanent, full time employee, thus we can maintain this service indefinitely. Servicing of the sensors (e.g., calibration, etc.) will be provided by CAWS or faculty research/course budgets. Faculty member Williamson and the associated students will provide basic deployment of the sensors as well as downloading of the data in the field.

Budget: Hardware

Hardware Title(s) & Vendor(s): Fondriest Inc., and Miami University Bookstore

Hardware Costs: $24,000.00

What is the total budget amount requested?: $24,000.00