Estimated tmax (way) too large for FourParam response function

[mbakker7]

tmax for the 4-parameter function needs more testing. It goes wrong, for example:

fourp = ps.FourParam()
fp = fourp.step([2, 1, 50, 100], dt=1, cutoff=0.95)

[raoulcollenteur]

Just to add: this means that tmax is too large, but it won't negatively affect models using this response function. You just convolve with a very long block response function.

[mbakker7]

Correct

…

On Thu, Jul 9, 2020 at 8:22 AM Raoul Collenteur ***@***.***> wrote: Just to add: this means that tmax is too large, but it won't negatively affect models using this response function. You just convolve with a very long block response function. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#209 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAKM3SEPJ34D5S3CZ3NASJTR2VOZDANCNFSM4OUVXQHA> .

[raoulcollenteur]

For future reference, here some code to test

cutoff = 0.999
fourp = ps.FourParam()
fp = fourp.step([2, 1, 50, 100], dt=1, cutoff=cutoff)
print(len(fp[fp<cutoff]))

returns 525 days. So after 525 days 99.9% of the response has passed. The length of fp is almost 1e4 days though.

So this issue is still relevant.

[mbakker7]

It may be possible to determine the cutoff on the fly when stepping through time in computing the step response.