Contents:

From: "Andrews Jim" <Andrews_Jim@msmail.middlebury.edu>

Date: 21 Mar 1996 10:15:26 -0500

Subject: Frequency of route visits

I am interested in how regularly the routes on the calling amphibian surveys have been run. If I understand correctly some designs call for only one survey during the calling period of each species. When I designed a protocol for a route here in Vermont my goal was to measure the changes in amphibian populations along that route. To do that, I felt that the route needed to be surveyed at least three times during the calling period of each species. I arrived at this protocol after I found out how easy it was to miss a given species when only a single survey was run along a route during a given breeding season. If the weather or listening conditions were not right on that one night it was very easy to miss a species. My protocol later evolved into three times per month in the months of April, May, June, and July. I don't believe that anything less than this (in northeastern US and Canada) will allow you to see changes along a given route. True when many routes are combined they may allow you to see changes in a state or a province, but you will not be able to locate specific sites from which a species has disappeared or appeared with any degree of confidence. I admit twelve times is a lot of times to ask a volunteer to go out with little or no notice. I had the advantage of paying the participants (no longer-no more funds). Although I suppose a single figure that indicates a decline in a state or a province is useful, it would be a lot more useful to know where the population changes were occurring so that we could start to figure out why. If a subset of the routes were run more frequently we could acquire some more useful data. I believe that conditions for calling and hearing differ so much even among wet nights, that missing the one or two prime nights will have a large impact on the data. Consequently I feel that along the intensively monitored routes it will be necessary to have a team of volunteers with an individual in charge of each route. If that person is going to be out of town for a period of time (or just at the movies or entertaining) they should designate another person to decide if the route should be run (in the monitoring where I did use volunteers I inevitably was left with gaps during crucial periods). Surveying routes once per breeding period will provide lots of good atlas data, perhaps a state-wide index for each species, and it definitely will be a good PR and conservation education tool, but I believe that a two-tiered system with intensive monitoring along some routes will be necessary to provide the kind of data that will allow us to answer some questions.

From FROG@nbs.gov Mon Mar 25 10:35:52 1996

Subject: Re: Frequency of route visits

An unknown person (unsigned message) wrote:

<snip>

"I arrived at this protocol after I found out how easy it was to miss a given species when only a single survey was run along a route during a given breeding season. "

<snip>

"True when many routes are combined they may allow you to see changes in a state or a province, but you will not be able to locate specific sites from which a species has disappeared or appeared with any degree of confidence."

<snip>

"Although I suppose a single figure that indicates a decline in a state or a province is useful, it would be a lot more useful to know where the population changes were occurring so that we could start to figure out why."

<snip>

It all depends on the question you are asking. The current protocol is designed to determine trends at the regional level and the author above correctly points out that you can safely state little at the level of the individual route (unless you have many, many years of data).

You can get more local information by mathematically smoothing the information among individual routes to create a contour map of increases and decreases. (You can see how this was done in the bird world by checking out the BBS home page via: ) However, while this would allow you to look at associations of decline within regions, it still tells you little about a single route.

You can test how effectively you monitor each species on a route with the data you collected. The question of whether a species is adequately monitored can be investigated by calculating the sampling variance from the runs you made and dumping that info into the MONITOR.EXE program. This program will tell you whether your current configuration (or any hypothetical configuration) has sufficient power to detect trends.

You can download that program from anonymous ftp site:

ftp.im.nbs.gov /pub/software/monitor

I keep harping on using this software because it is only recently that such a program has existed and so many monitoring programs are started without any statistical reality check.

sam

   0<    0-      0<    0-      0<    0-      0<
->>)   ->>)   ->>)   ->>)   ->>)   ->>)   ->>) 
  L      L      L      L      L      L      L

Sam Droege FROG@NBS.GOV

w 301-497-5840 h 410-798-6759 fax 301-497-5784

NBS, 12100 Beech Forest Dr., Laurel, Md 20708

/im.html

One is for bad news

Two is for mirth

Three is a wedding

Four is a birth

Five is for riches

Six is a thief

Seven is a journey

Eight is for grief

Nine is a secret

Ten is for sorrow

Eleven is for love

Twelve is for joy tomorrow

- traditional

From: Paul Geissler, not submitted to discussion group

I agree with Jim that "the route needed to be surveyed at least three times during the calling period of each species" and "that anything less than this . . . will allow you to see changes along a given route." However, the best way to detect changes in a State, Province, or other large area, is to survey a route only once during the calling period of each species and to add more routes. Usually you can survey a new route almost as easily as resurveying the same route. But you get less information when you resurvey a route because you are recounting the same individual frogs. In other words, there is a correlation among the surveys on the same route, reducing the new information content of the data. Technically, there is a sampling error component of variance associated with routes and a measurement error component associated with resurveying the same route. Surveying a new route reduces both components of variance, but resurveying a route only reduces the latter component.