I use GRADISTAT.xls grain size analysis for marine sediments. The sieve sizes that I use are different from those given in the excel. For example the excel has 4000 then 2800 and 2000. I do not have a value for 2800. Is it better to put value 0 into this sieve or delete it and continue with 2000 sieve? Because I did it both ways and the results are different! Which one is the correct one? Or is it another way?
Thank you!
Hi,
Base on how you describe it, I would remove the "0" class as this is not a size fraction, especially if it is coming in the middle of 2000 and 4000 um.
You should include the sieve sizes and ensure you are completing all other required fields and everything should be ok!
Take care in using GRADISTAT in case you want Falk & Ward 1957 parameters. G makes linear interpolation between points (sieve size, %), but F&W method is based on a probabilistic distribution. I.e. straight lines can connect point only if you have a probabilistic scale. If you use a linear scale, your error will be on average 1/2 size interval. This means that if you sieve at 1/2 phi, your error is 1/4 phi. In these conditions, a minor issue is the lack of a sieve and how you solve the problem. I suggest graphically reconstructing that part of the distribution on a probabilistic scale and give the resulting value to the lacking class. Some softwers have probabilistic distribution, e.g. Origin.
- Badges
- Science topic