Owl team response to questions from Steve Brink

Owl team response to questions from Steve Brink by Kim_Ingram, at 11:36 a.m. on 6 September 2012,

The following response is from the SNAMP Owl team to questions submitted by Steve Brink after the IT meeting of 8/23/2012:

Q1:Looking again at the Lambda graph, it’s interesting that the downward slope of the line is nearly constant. I should have thought about the 2001 Star Fire when Pat Ferrell, Eldorado NF Contracting Officer, asked a question about it. The Star Fire burned 16,900 acres in Aug-Sept. 2001, about 2,400 acres of which were on the Georgetown Ranger District. Following the fire, the USFS concluded that the two HRCAS, in which PAC055 and PAC075 were located, were essentially destroyed and recommended deleting both PACs from the Forest plan. Further, about 70% of the total acres burned in the Fire were of high severity. So, how is it that the Star Fire had absolutely no affect on the owls?

Response: First, recall that Doug noted that the estimates for lambda are based on model averaging (the average estimate of several models), which results in the smooth (“constant”) curve to which you refer. As long as the declining trend is not abrupt, it will appear relatively smooth despite slight variation in the annual estimates. Second, the two owl territories affected by the Star Fire were Regional territories, and the lambda analysis only included owl territories on the Density Study Area. Even if the two owl territories affected by the fire had been located on the Density Study Area, the observed impact would not have been dramatic because there are 45 owl territories on the density study area. In addition, we note that about 2,500 acres did burn on the Eldorado Density Study area during the Star Fire. This burn affected 3 owl territories on the density study, but all have been occupied at least once since the fire.

Q2: Because soil site classification could be used as a surrogate for the overall “primary productivity” of a forest, will the samples for the control and “treatment” areas be stratified? The premise being that owls in highly productive forests will perform better under both the control and “treatment” conditions, and likewise owls on less productive forests will underperform under control or treatment scenarios.

Response: This was suggested in the context of including “abiotic” factors as covariates in our analysis, which we think is a good idea. The idea of including soil type is a good one and we will consider that thoroughly for our list of covariates.

Q3: It’s still unclear to me why the Team is not opportunistic and looking to tease out why Owls had very high reproductive success in 1992. Using the 1993 data, and not using 1992, doesn’t seem intuitive given Rocky’s premise that CSO’s live a long time and have an evolutionary strategy that is opportunistic for reproduction. How can that be justified?

Response: The owl team has often wondered why there was such high reproduction in that year. The problem is that there is that there have been no other years such as 1992 to make comparisons. In other words, if you found several factors that were different in this year relative to other years, how would you distinguish which was consistent in terms of its importance to owls? One needs to have multiple examples of unique events to detect a pattern. In addition, 1993 was the first year that we had good aerial photography for the entire study area and adequate sampling effort to have confidence in detecting reproduction and a sufficient number of surveys over the entire study area. We note also, that unlike what was implied in our presentation, we do have air photos prior to 1993 but they are not seamless digitized photos as in 1993. Thus, it is quite time consuming to work with these other photos. It is important for the record to note that all of the effort to create vegetation maps by the owl team for the past dozen years in order to assess impacts has been done, until last year, without funding. That is, we have no allocated money to engage in this activity, but all the owl investigators have devoted their own time to create such maps because we realize the importance of the information. We continue to assess sources of air photographs that are easily used and welcome any help the SNAMP stakeholders are able to offer.

Q4: Given how robust the IPM is, why wouldn’t Doug include ALL of the owl data that the Team has including: 1)The discrepancies related to the 8 birds that were counted in the occupancy but not as part of the population and likewise the other 4 birds that resisted capture? 2)The 1992 reproduction data since it is congruent with Rocky’s premise that the evolutionary strategy of the owl is to be long lived and be opportunistic for reproduction? 3)Running a comparison with and without the 1992 data?

Response: “All” of the birds, including the ones referred to in part 1 of this question, are incorporated into the IPM through the count data. The count data are the number of birds detected each year on the Density Study Area, regardless of whether the birds are marked or unmarked. Doug chose 1993 as the starting year for all of his population trend analyses because that was the first year that we achieved adequate survey coverage of the Density Study Area. As Doug discussed, it is critical that the study area be consistent in size when assessing population trends because the number of birds in the population is a direct function of the size of the study area. The 1992 data are consistent with the hypothesis that spotted owls have evolved a “bet hedger” life history strategy, but so are many other features of their life history (such as long life span, low fecundity). So the 1992 data are interesting but they do not make or break any analysis we do. They are interesting and we are curious, like you, as to why we have not seen similarly high reproduction years since then. However, these data as a single event do not provide insight into changes in owl populations other than to demonstrate that unusually high years of reproduction do occur infrequently.

This post is a part of the following discussions: