Work Description

Title: Mono Lake Water Levels Forecasting using Machine Learning Open Access Deposited

h
Attribute Value
Methodology
  • This thesis develops three machine learning approaches to forecast water levels at Mono Lake, California, each approach utilizing distinct datasets as outlined below and depicted in Figure 1 of the thesis. I. Using historical water levels dataset II. Using historical precipitation dataset III. Using historical Bayesian statistical dataset The first approach used the historical water levels dataset which consists of the historical water levels of Mono Lake measured in feet (Ft). The water levels dataset has a monthly frequency spanning from 1/4/1941 to 12/1/2018. This dataset was sourced from the Mono Lake Committee website (Mono Lake Levels 1979-Present (Monthly)). The historical water levels are used to develop machine learning models. The second approach used the historical precipitation dataset, specifically from ERA6 (Hersbach et al.; Lavers et al., 2022), CRUTS_adj (NOAA Physical Sciences Laboratory, n.d.), and MERRA3 (MERRA, NASA). The precipitation datasets have a monthly frequency spanning from 1/1/1980 to 12/1/2021 and are measured in inches. Hydrology experts have extrapolated these precipitation global datasets for our study area for the Mono Lake basin (Gossard et al., 2023). We then utilized a Bayesian model to combine these three datasets (ERA6, CRUTS_adj, MERRA3). The output from these Bayesian models is subsequently used to develop machine learning models. For our third approach, we used the historical Bayesian statistical dataset, which is the output from the Large Lake Statistical Water Balance Model (L2SWBM) that includes precipitation, evaporation, and runoff datasets. L2SWBM also has used the global datasets that have been extrapolated for the Mono Lake basin (Gossard et al., 2023). For this research, we have selected the median values, produced by the L2SWBM to minimize skew from extreme data points. The timeframe for this dataset spans from 1/1/1970 to 12/1/2018 with data measured in thousands of acres-feet (kAc-Ft) of water. The supplementary material includes stage-curve calculations using Smoothed Pelagos Corporation Bathymetry Data of Mono Lake as detailed in thesis Appendix I (Table A-1. Bathymetry of Mono Lake, n.d.). Specifically, using the Bathymetry data we converted the observed water levels into change in storage. The data for Mono Lake water levels was collected from the Mono Lake Committee website (Mono Lake Levels 1979-Present (Monthly)). In this calculation, we used stage-volume linear interpolation to calculate the volume of water in Mono Lake corresponding to the observed water level and subsequently calculated the change in storage for each month. Finally, with the calculated change in storage and the precipitation, evaporation, and natural runoff from the L2SWBM, the only remaining component in the Water Balance Equation (Equation 1 in the thesis) is the diversions. Here we used the water balance equation to calculate the diversions per month.
Description
  • The three approaches used three distinct datasets named as follows: Historicalwater_levels.csv, Historical_Precipitation.csv, and Bayesian Statistical dataset.csv. These files are accessible using Microsoft Office or similar software. The machine learning models are developed in Jupyter Notebook (.ipynb) files, named according to the datasets they utilize. However, for the third approach, the models are named Random Forest, LSTM Model Base, and Multivariate LSTM Models. More details are available on the Shah_Bhavarth_Readme.txt. These notebooks can be accessed through Python, Project Jupyter, or Google Colab, and dependencies include libraries such as Pandas, NumPy, Matplotlib, Scikit-learn, Keras, and TensorFlow. The supplementary material also includes Excel files for stage-curve calculations and diversions, named Water_levels_Stage_Curve_Calculations1970-2018.xlsx and Diversions_calculation.xlsx, respectively.
Creator
Creator ORCID
Depositor
  • bhavarth@umich.edu
Contact information
Discipline
Keyword
Citations to related material
  • Shah, Bhavarth. 2024. "Mono Lake Water Levels Forecasting Using Machine Learning." Master’s thesis, University of Michigan, School for Environment and Sustainability. ORCID iD: 0000-0002-2391-8610. https://dx.doi.org/10.7302/22659
Related items in Deep Blue Documents
Resource type
Last modified
  • 05/03/2024
Published
  • 04/29/2024
Language
DOI
  • https://doi.org/10.7302/xbet-5212
License
To Cite this Work:
Shah, B. (2024). Mono Lake Water Levels Forecasting using Machine Learning [Data set], University of Michigan - Deep Blue Data. https://doi.org/10.7302/xbet-5212

Relationships

This work is not a member of any user collections.

Files (Count: 26; Size: 28.9 MB)

Thumbnailthumbnail-column Title Original Upload Last Modified File Size Access Actions
Shah_Bhavarth_Readme.txt 2024-04-23 2024-04-29 3.24 KB Open Access
Historicalwater_levels.csv 2024-04-23 2024-04-23 16.2 KB Open Access
Water_Levels_Forecast_Support_Ve...ipynb 2024-04-23 2024-04-23 122 KB Open Access
Water_Levels_Forecast_Random_For...ipynb 2024-04-23 2024-04-23 133 KB Open Access
Water_Levels_Forecast_LSTM_Model.ipynb 2024-04-23 2024-04-23 164 KB Open Access
Historical_Precipitation.csv 2024-04-23 2024-04-23 23.3 KB Open Access
Simple_Bayesian_Based_Precipitat...ipynb 2024-04-23 2024-04-23 266 KB Open Access
Advanced_Bayesian_Based_Precipit...ipynb 2024-04-23 2024-04-23 266 KB Open Access
Bayesian_Statistical_dataset.csv 2024-04-23 2024-04-23 40.3 KB Open Access
Random_Forest_Model.ipynb 2024-04-23 2024-04-23 1.09 MB Open Access
Random_Forest_Model_Tuning.ipynb 2024-04-23 2024-04-23 1.09 MB Open Access
LSTM_Model_Base.ipynb 2024-04-23 2024-04-23 1.75 MB Open Access
Multivariate_LSTM_Model_1.ipynb 2024-04-23 2024-04-23 1.73 MB Open Access
Multivariate_LSTM_Model_2.ipynb 2024-04-23 2024-04-23 1.77 MB Open Access
Multivariate_LSTM_Model_3.ipynb 2024-04-23 2024-04-23 1.74 MB Open Access
Multivariate_LSTM_Model_4.ipynb 2024-04-23 2024-04-23 1.76 MB Open Access
Multivariate_LSTM_Model_5.ipynb 2024-04-23 2024-04-23 1.75 MB Open Access
Multivariate_LSTM_Model_6.ipynb 2024-04-23 2024-04-23 1.65 MB Open Access
Multivariate_LSTM_Model_7.ipynb 2024-04-23 2024-04-23 1.77 MB Open Access
Multivariate_LSTM_Model_8.ipynb 2024-04-23 2024-04-23 1.78 MB Open Access
Multivariate_LSTM_Model_9.ipynb 2024-04-23 2024-04-23 1.77 MB Open Access
Multivariate_LSTM_Model_10.ipynb 2024-04-23 2024-04-23 1.74 MB Open Access
Multivariate_LSTM_Model_11.ipynb 2024-04-23 2024-04-23 1.76 MB Open Access
Multivariate_LSTM_Model_12.ipynb 2024-04-23 2024-04-23 1.76 MB Open Access
Diversions_calculation.xlsx 2024-04-23 2024-04-23 59.9 KB Open Access
Water_levels_Stage_Curve_Calcula....xlsx 2024-04-23 2024-04-29 2.96 MB Open Access

Date: 23 April, 2024

Dataset Title: Mono Lake Water Levels Forecasting using Machine Learning

Dataset Creators: Shah, Bhavarth

Dataset Contact: Bhavarth Shah bhavarth@umich.edu

Key Points:
- As outlined in the main thesis, we have developed three approaches to forecasting water levels at Mono Lake, California using machine learning. Each approach utilizes a different dataset.
- These approaches include (Figure 1 in main thesis):
I. Using historical water levels dataset
II. Using historical precipitation dataset
III. Using historical Bayesian statistical dataset

Data Description:
- The three approaches used three distinct datasets named as follows: Historicalwater_levels.csv, Historical_Precipitation.csv, and Bayesian Statistical dataset.csv. These files are accessible using Microsoft Office or similar software.

- The supplementary material includes Jupyter Notebooks (.ipynb) and Microsoft CSV and Excel (.csv and .xlsx) files. These notebooks can be accessed through Python, Project Jupyter, or Google Colab, and dependencies include libraries such as Pandas, NumPy, Matplotlib, Scikit-learn, Keras, and TensorFlow. The CSV and Excel files can be opened with Microsoft Excel, Google Sheets, or similar applications.

- The supplementary material also includes Excel files for stage-curve calculations and diversions, named Water_levels_Stage_Curve_Calculations1970-2018.xlsx and Diversions_calculation.xlsx, respectively.

Files contained here:
The files are divided based on the above three approaches. The corresponding CSV and ipynb files correspond to the respective dataset (.csv) used and the machine learning model (.ipynb) developed using that dataset.

# I. Using historical water levels dataset
- Historicalwater_levels.csv
- Water_Levels_Forecast_Support Vector Machine Model.ipynb
- Water_Levels_Forecast_Random Forest Model.ipynb
- Water_Levels_Forecast_LSTM Model.ipynb

# II. Using historical water levels dataset
- Historical_Precipitation.csv
- Simple_Bayesian_Based_Precipitation_Forecasting.ipynb
- Advanced_Bayesian_Based_Precipitation_Forecasting.ipynb

# III. Using historical Bayesian statistical dataset
- Bayesian Statistical dataset.csv
- Random Forest_Model.ipynb
- Random Forest_Model_Tuning.ipynb
- LSTM_Model_Base.ipynb
- Multivariate LSTM_Model_1.ipynb
- Multivariate LSTM_Model_2.ipynb
- Multivariate LSTM_Model_3.ipynb
- Multivariate LSTM_Model_4.ipynb
- Multivariate LSTM_Model_5.ipynb
- Multivariate LSTM_Model_6.ipynb
- Multivariate LSTM_Model_7.ipynb
- Multivariate LSTM_Model_8.ipynb
- Multivariate LSTM_Model_9.ipynb
- Multivariate LSTM_Model_10.ipynb
- Multivariate LSTM_Model_11.ipynb
- Multivariate LSTM_Model_12.ipynb

# Calculations and conversions:
- Diversions calculation.xlsx
- Water levels_Stage_Curve_Calculations1970-2018.xlsx

# Original Thesis:
- Shah_Bhavarth_Thesis.pdf

Please note that the CSV files referenced in the machine learning model (.ipynb) files may have names that differ from those listed above. However, the data within these CSV files remains the same.

Use and Access:
Attribution-NonCommercial 4.0 International (CC BY-NC 4.0).
This license lets others modify and build upon your work non-commercially. Their new works must also acknowledge you and be non-commercial, but they don’t have to license their derivative works on the same terms.

To Cite Data:
Shah, Bhavarth. 2024. "Mono Lake Water Levels Forecasting Using Machine Learning." Master’s thesis, University of Michigan, School for Environment and Sustainability. ORCID iD: 0000-0002-2391-8610.

Download All Files (To download individual files, select them in the “Files” panel above)

Best for data sets < 3 GB. Downloads all files plus metadata into a zip file.



Best for data sets > 3 GB. Globus is the platform Deep Blue Data uses to make large data sets available.   More about Globus

Remediation of Harmful Language

The University of Michigan Library aims to describe library materials in a way that respects the people and communities who create, use, and are represented in our collections. Report harmful or offensive language in catalog records, finding aids, or elsewhere in our collections anonymously through our metadata feedback form. More information at Remediation of Harmful Language.