[Eoas-seminar] EOAS Colloquium Friday Jan 20 at 3pm

[eoas-seminar at lists.fsu.edu]

Please join us for this week's EOAS Colloquium speaker Friday Jan 20 at 
3pm in 1050:


*Model- and Data-Based Approaches for Identifying Controlling Processes 
of an Environmental System and my View of Environmental Data Science*

**

Dr. Ming Ye (mye at fsu.edu <mailto:mye at fsu.edu>), Professor

Department of Earth, Ocean, and Atmospheric Science and Department of 
Scientific Computing

Florida State University

An environmental system is open and complex, and its evolution in space 
and time involves a large number of natural and anthropogenic processes 
and their interactions. It is necessary to identify controlling 
processes and process interactions for advancing our understanding of 
the environmental system to support science-informed decision-making on 
environmental protection. This seminar presents both model- and 
data-based approaches that we have recently developed for identifying 
controlling processes. The model-based approaches use global sensitivity 
analyses, and have a unique feature of addressing uncertainty not only 
in process representations (i.e., process models) but also in process 
parameters (i.e., process model parameters). The approaches are designed 
to answer the following question: if we are not certain about process 
representations and process parameter values, can we identify important 
or influential processes for developing and improving environmental 
system models? The model-based approaches are always computational 
expensive, but can be applied to an environmental system with scarce 
data. The data-based approaches do not require developing process 
models, but use data to identify spatio-temporal patterns and the 
processes that control the patterns. This is achieved in our studies by 
using cluster analysis (one-way clustering and co-clustering) and 
/t/-distributed Stochastic Neighbor Embedding (t-SNE) methods. An 
example application will be presented for a dataset of 13,024 
groundwater geochemical measurements for 11 geochemical parameters of 
1,184 groundwater samples collected over 23 years from 29 monitoring in 
a regional aquifer. The model- and data-based approaches are an 
integration of environmental science, mathematics and statistics, and 
computer science. At the last of the presentation, I will discuss our 
research on study sinkhole lakes at Lake Miccosukee and Lake Jackson 
located in Tallahassee.


*The two figures below are for illustrating the model- and data-based 
approaches. *


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.fsu.edu/pipermail/eoas-seminar/attachments/20230116/6050eadc/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: GFKetAk4mz0nNj3g.png
Type: image/png
Size: 116635 bytes
Desc: not available
URL: <http://lists.fsu.edu/pipermail/eoas-seminar/attachments/20230116/6050eadc/attachment.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 7oeNrOK5EB4VLvJy.png
Type: image/png
Size: 752444 bytes
Desc: not available
URL: <http://lists.fsu.edu/pipermail/eoas-seminar/attachments/20230116/6050eadc/attachment-0001.png>