Channel banner for maca
Channel avatar for maca
Support climate action at all levels.

Report Postclose

Remove Postclose

Are you sure? After you remove the post, it will no longer appear in channel listings but you can access it directly. You can undo this later by clicking "approve".

Delete Postclose

Are you sure you want to delete this post? This is a permanent action and cannot be reversed.

Machine Learning, Biochar, and Cattle Ranching
Profile image for Chris Harding
Chris Harding
 — Chemical Engineer and Biological Scientist
10 months ago

I learned from suggesting the wrong pathway to my Board of Directors. 

I thought of AI instead of computational fluid and combustion dynamics but I found it more daunting. I was afraid I couldn't do it. So, I was anxious and negative. John nudged me in the right direction. 

I spent the day researching AI's relationship to biochar production. There is a lot of data available. The first I started with was a mini-review of Machine Learning (ML) methods for different predictions of biochar properties like yield, specific surface area properties, cation exchange properties, etc. I then bought two excellent books on advanced ML and AI. I read the first chapter. I can do the math, with some learning, but I don't have to do the math. All I have to do is MATLAB and Python programming with an understanding of the algorithms. 

I have mentioned the USDA NCRS Web Soil Survey for soil data, but I found https://www.pnwbiochar.org/ as well. It uses the physical and chemical properties of soil and crops and suggests biochar, biomass, and tons/Acre of biochar. It also provides a rationale for large or small particles of biochar if water runoff needs to happen or not happen, respectively.  

I discovered a University of California and Oregon State journal article about using biochar in pastures[1]. I had looked for the article because Oregon has more pastures than farmland. Also, Discernable Spectrum can likely optimize our ML methods on increased grass growth without a worry for catastrophic destruction of a section of crops--I don't think that would happen, but farmers might think that. So, there is a cattle rancher in my hometown. I will present a summary of [1] and collect more research articles after I learn AI, and practice on the potential data from [2] if they decide to share it--they said to make a request. I have already evaluated his land with Web Soil Survey and it has an average pH of 4.75--different in different areas. This is a little low and the range was pH 4.3 to 5.2--there are 4 unequal sized areas in the cattle ranch. It also has a high Organic Carbon content average of 14.75% with a range of 6% to 33.9%. So, biochar with ash is needed. The above link suggested Oregon Redwood at 700 degrees C. This is good because a machine learning paper [2] showed that biochar yield increases with ash and temperature. Also, the feed stream can be oxygenated so regular air will work if a fluidized bed is used. Since there is high organic carbon, basalt would be useful because minerals help increase the fertility of high organic carbon[see link above]. 

Anyhow, I will have a renewed action plan by next Monday. I need to complete the Mission, Vision, and Values statements first. Now, we have cattle land and farmland to consider. Note, I know beef is bad for the environment, but I doubt people give it up any time soon so we might as well capture N2O and CO2 and reduce fertilizer use. 

References: 

1. Gao, S., & DeLuca, T. H. (2022). Rangeland application of biochar and rotational grazing interact to influence soil and plant nutrient dynamics. Geoderma, 408, 115572. https://doi.org/https://doi.org/10.1016/j.geoderma.2021.115572


2. Haq, Z. U., Ullah, H., Khan, M. N. A., Raza Naqvi, S., Ahad, A., & Amin, N. A. S. (2022). Comparative study of machine learning methods integrated with genetic algorithm and particle swarm optimization for bio-char yield prediction. Bioresource Technology, 363, 128008. https://doi.org/https://doi.org/10.1016/j.biortech.2022.128008

arrow_upward1
rss_feedFollow

Delete Postclose

Are you sure you want to delete this post?