Standardizing data
Bruce, Standardizing might not be the best way to go if you have low counts. You can possibly assume that events follow a homogeneous Poisson process and rate varies with night length (linear or quadratic) [Y|x ~ Poisson(phi); log(phi)=f(x)]. You can estimate corresponding coefficients by glm(). I think controlling for night length differences work even if you throw in other covariates as additive effects on the log scale. You can get sunrise /sunset times from maptools::sunriset, be careful with timezones though. Cheers, Peter -- P?ter S?lymos, Dept Biol Sci, Univ Alberta, T6G 2E9, Canada AB solymos at ualberta.ca, Ph 780.492.8534, http://psolymos.github.com Alberta Biodiversity Monitoring Institute, http://www.abmi.ca Boreal Avian Modelling Project, http://www.borealbirds.ca
On Wed, Jan 23, 2013 at 7:16 PM, Bruce Miller <batsncats at gmail.com> wrote:
Hi all,
Posting this query again as no one replied.
I need to create a way to standardize bat sampling data in northern
latitudes not only by the Activity Index standardized by sample time
(unit effort) but by the constantly changing night length.
This was easy in the tropics with a straight forward application of my
Acoustic Activity Index (AI) [Miller, B. W. 2001. A method for
determining relative activity of free flying bats using a new activity
index for acoustic monitoring. Acta Chiropterologica. 3: 93-105].
This uses the occurrence of each species by 1-minute time blocks and is
then standardized by "Unit Effort" which is the total sample time each
night in hours. With a night length that was fairly standard with only a
small ? of 1 hr or so for seasonal variation and very small
night-to-night change this seem OK.
However looking at data from here in the northern latitudes,I need
something else in addition. The night length not only changes rapidly
night-to-night during the summer, but has a very wide ?. So using only
AI -occurrences per each 1-minute time interval standardized by sampling
time previously used in the tropics may not reflect a realistic
comparison measure up here in the north.
So my question is how to incorporate the length of each night into the
AI standardized by sample time?
How best to integrate the night length since that is one of the key
factors constraining bat activity?
*/R/* seems awesome for running repetitive calculations once one has the
code line script. So I am trying to see how to develop such a new
standardized /R/ code for two data sets, one a DF with the simple AI
values that have already been standardized by unit effort and another
that includes the ever changing sunset-sunrise data.
For working out GGplot temporal activity plots by minute for each night
it took bit of hand holding by Hadley as R does not (or at lead did not)
"do time well". Not sure what can be done with standardizing the data
with changing night lengths form one night to the. The crossover
midnight may not be an issue for this calculation since we just need a
total night length in decimal hours.
Moon phase and % illumination is yet another issue, but not relevant at
the moment.:-)
Any and all suggestions welcome and the bats will depend on it.
Bruce
--
Bruce W. Miller, Ph.D.
Conservation Ecologist
Neotropical Bat Projects
office details
11384 Alpine Road
Stanwood, Mi. 49346
Phone (231) 679-6059
[[alternative HTML version deleted]]
_______________________________________________ R-sig-ecology mailing list R-sig-ecology at r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-ecology