Manuel--
I apologize in advance for not answering the exact question you ask about
packages. [It is included in some geostatistics packages in terms of
semivariance, nugget, sill, etc.]
In ecological data, time to independence is very scale dependent. There's
autocorrelation at scales of seconds due to instrument
temperature-dependence if that hasn't been calibrated for, or the same
individuals in the camera trap frame. That component of dependence may
have a half-life of minutes. There's often autocorrelation based on time
of day & temperature, with cycles of 24 hours. There may be pulse events
from storms that persist a few days. There's seasonality driving
temperatures, day lengths, and plant & animal behavior, with cycles of 1
year. Then where I live there are ENSO-driven temporal dependence at
scales of 1.5 - 3 years, PDO at about a decade, and ENSO-La Nina dominated
periods of 4-6 decades that drive not just ocean ecology, but rainfall &
thus terrestrial ecology. Then there's tends up to climate change.
So, in my experience in optimizing sampling designs for monitoring for
trends, the majority of the temporal dependence is driven by cycles or
pulses of characteristic duration, and that is more useful for determining
the sampling frequency than empirical estimation form a "continuous"
datastream of limited duration. That approach also helps me think about
the spatial concordance of the correlated errors: which are site-specific,
which are concordant across all of the sites.
Tom
-----Original Message-----
From: R-sig-ecology <r-sig-ecology-bounces at r-project.org> On Behalf Of
Manuel Sp?nola
Sent: Wednesday, March 3, 2021 1:06 PM
To: r-sig-ecology at r-project.org
Subject: [EXTERNAL] [R-sig-eco] Time to Independence in R
This email has been received from outside of DOI - Use caution before
clicking on links, opening attachments, or responding.
Dear list members,
It is common in ecology to sampling in almost a continuous manner when
using data loggers, camera traps, sound recorders, gps radio-collars. etc.
Is there any R package to assess time to independence for the data to
avoid temporal autocorrelation?
I know that there are models to take into account the temporal
autocorrelation of the data, but I am asking to optimize the data
collection, before modeling.
Thank you very much in advance.
Manuel
--
*Manuel Sp?nola, Ph.D.*
Instituto Internacional en Conservaci?n y Manejo de Vida Silvestre
Universidad Nacional Apartado 1350-3000 Heredia COSTA RICA mspinola at una.cr
<mspinola at una.ac.cr> mspinola10 at gmail.com
Tel?fono: (506) 8706 - 4662
Personal website: Lobito de r?o <
https://gcc02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fsites.google.com%2Fsite%2Flobitoderio%2F&data=04%7C01%7Ctom_philippi%40nps.gov%7C7c92bae61c0d458d045908d8de8881b0%7C0693b5ba4b184d7b9341f32f400a5494%7C0%7C0%7C637504025171590082%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=qTaPgnDdCNuCtww7koY%2F7StrjYRczIa2YNZiNveJo%2FQ%3D&reserved=0
[[alternative HTML version deleted]]