I believe that most or all researchers involved with the demographic studies agree that their populations are, in fact, declining, but we cannot be certain that they are declining as steeply as the l values indicate. In our studies in the Sierra National Forest (SNF) and Sequoia/Kings Canyon National Parks (SNPs), for example, the rate of decline among territorial birds, by actual count, is less than that suggested by the l values.

The general conditions and loss of older forests since European settlement times is speculative in relation to the current owl population in the Sierra Nevada. It remains a plausible, but unsubstantiated, hypothesis. We note, for example, that spotted owls are not confined to the few remnant stands of old-growth forests in the Sierra Nevada. Rather, they occur in a wide range of different habitats so long as those include relictual attributes of old-growth, particularly the very old and decadent trees that provide suitable nest sites. Indeed, given the uncertain nature of the suitability for owls of the open, parklike structure of the forests that existed over much of the early Sierra Nevada, the California Spotted Owl Technical Assessment Team stated in the CASPO Report that "Because we lack adequate, historical inventories of spotted owls in the Sierra Nevada, we have no basis for comparison with our current knowledge. Their current distribution and abundance, however, do not suggest that they have declined either in their overall distribution in the Sierra Nevada or that they have declined markedly in abundance within any forest type" (Verner et al. 1992 @ 11). Further, "Indeed, spotted owls may be more abundant in some areas of the Sierra Nevada today than they were 100 years ago" (Verner et al. 1992 @18).

Several studies of factors associated with reproductive performance of spotted owls in California have now shown that weather is a major player, especially late winter and spring storms. In our demographic study areas in the SNF and SNPs, we have repeatedly noted the failure of nesting owls as a result of heavy storms in March, April, May, and even June--this year 5 inches of snow fell over our high-country study sites on 2 June (my backyard in Fresno was blanked by about an inch of large hail stones that essentially obscured the lawn!). As a result, five of seven active owl nesting attempts above 5,500 feet elevation in the SNF study area failed. Where only light rain fell, between 3,400 and 5,200 feet in elevation, only one of eight active nests failed. The fact that the owls initiate breeding in any given year suggests that their response to general habitat conditions is positive, that their response to prey availability is positive, but that they are probably even less adept than the National Weather Service at forecasting the weather.

To state emphatically that recently observed population declines of the spotted owl are "resulting from loss, degradation and fragmentation of habitat," without due consideration for various other possible contributory factors, such as fluctuations in prey availability and the widely observed, clearly established, and catastrophic effects of severe weather on owl breeding success, runs counter to well-established standards of scientific inquiry.

In a 1992 report from the California Forestry Association, Robert J. Taylor reported the following: "At this time 17.9% of industrial forest lands have been surveyed... surveyors have discovered 135 owl sites...many surveys not yet complete...this number could either rise...or fall... If the surveyed lands are typical of all industrial lands, then the estimated total of owl territory equivalents is 754... Industrial forests appear to support 42% more owl sites per square mile than suitable habitat on the national forests. The forest Service estimate is below the lower bound of the error interval for industry lands...[but]...is within the range of densities for the demographic study sites (CASPO Chap 8)..." This report by Taylor was cited in the CASPO Report, along with a section in the report (@ 87) that discussed owl nest sites in private timberlands, so the persons who prepared the Status Report could have known to obtain a copy if they didn't have one.

From the above, it should be clear that the true number of spotted owl sites on private lands in the Sierra Nevada is not well known, so my point here is not to try to establish that number. Instead, my point is to plea for an even hand when selecting facts to make a point. Both of these sources suggest more owls and owl sites on private timberlands than readers would conclude from the Status Report alone.

The Taylor report cited above also discussed nesting by spotted owls solely on Fibreboard Industries lands and a mix of their lands with Forest Service lands. Taylor's table 5 (@ 19) summarizes results over 1990, 1991, and 1992, showing a total of 60 sites surveyed for nests and 33 of those (55%) confirmed for reproduction. Taylor states (@ 8) "A total of 22 pairs nesting on industry lands were surveyed for success at fledging young. These produced 41 fledglings, for a mean productivity of 1.864 young per nest. If Fibreboard's percentage of pairs breeding is typical of industry lands, then fecundity has averaged 1.03 young of both sexes per pair. This is equal to the performance of the Sierra National Forest, the location showing the highest fecundity of any of the demographic study areas (CASPO Chap 8)." We might disagree with Taylor's extrapolation to all commercial timberlands in the Sierra Nevada, but my point here is that data exist to show that owls can and do nest on commercial forestlands. Uncertainty remains, at least for the general public, about how prevalent and successful these efforts are. But to cite only the data of Bias and GutiÈrrez (1992) as evidence of what is occurring on commercial forests in the Sierra Nevada misrepresents the true situation.

Just for clarification, the l value for SNF was 0.898, or essentially 0.90.

It isn't clear to me how these conclusions were reached, given available data. The annual report for 1998 of the SNF and SNPs studies (Steger et al. 1998) does not state whether the difference in l between the two study areas is statistically significant. At my request, Dr. James Baldwin, PhD biometrician at our Station's headquarters in Albany, ran a two-sample z-test on the two l values and concluded that the z score is far less than the value required for significance at the 0.05 level.

Next, Table 1 (@ 4) of the Status Report purports to show three significant differences between annual tallies of owls in the demographic study areas in the SNF and SNPs over the 9 years from 1990 to 1998. I consulted with Sylvia Moria, one of our Station's biometricians, about this analysis, and she expressed doubt that it was a valid analysis. I then FAXed copies of pages 4 and 5 of the Status Report to her, and she was unable to duplicate the analysis. At this point, I requested a full biometrics review of the Status Report, the results of which I will need to report later.

Perhaps the differences between numbers of birds detected in the two study areas led writers of the Status Report to conclude that the l values were significantly different. If so, this is not an appropriate way to make that determination. Consider an example of two populations, one with 5,000 individuals and one with only 100 individuals. If both have l values of 0.95, both are declining at the same rate--they do not differ significantly in this respect, even though they certainly differ significantly in numbers of individuals. If the rates of decline are different, however, they would have different trajectories (slopes), and one could test the significance of this difference by linear regression. With this in mind, visual inspection of Figure 1 (@ 5) in the Status Report clearly does not support a conclusion that the owls are faring better in SNPs. Indeed, it looks instead as though the numbers of individuals are currently declining at about the same rate. If this is the case, it argues strongly against a habitat basis for the difference between SNF and SNPs, thus implicating some other environmental factor. This was the immediate and unanimous response of three or four spotted owl scientists when we first viewed the Status Report during a break at a workshop in Sacramento on 23 June. Weather is the only clearly evident factor to surface to date. Linear regression of the two declining trends from 1995 to 1998 indicates a decline of 5.31% per year in SNPs and 4.36% per year in SNF; the slopes of the two lines are not significantly different, nor are their y-intercepts. The main difference is that the SNPs population reached peak numbers a year after SNF (in both cases undoubtedly related to the extraordinarily successful reproduction in 1992), so the decline in SNPs was delayed by one year. Why this was true is unknown, but it probably had something to do with the tendency for adults to hold territories longer in SNPs than in SNF.

The current field season is well underway, in fact far enough along to provide a fair indication of the numbers of owls present in these study areas this year. The count in the SNF study is presently at 54 birds, up four (8%) from 1998; the count in the "New Sierra" study area is at 22, down two from 1998; and the count for SNPs is at 49 birds, down nine (15.5%) from 1998. George Steger has informed me that he expects the total count to increase slightly in all areas before the field season ends, but probably not enough in SNPs to show an increase over 1998. As with so many other data currently available from our demographic studies, it's simply too early in the game to understand the significance, if any, of this new information.

Comparing densities of owls between SNF and SNPs is problematic at best. In any case, an attempt to estimate "ecological density" would be preferred over the estimate of "crude density" used in our annual reports to date. First, all three of our study areas have high edge-to-area ratios, complicating decisions about how to deal with territories near study-area boundaries. Then one would need to adjust for (i.e., take out) those portions of the two study areas (SNF and SNPs) that are clearly not suitable habitat. Even with all of that, the data would still suffer from uncertainties about how well the field crews managed to detect all birds in each area, or at least an equal proportion of all birds in each area.

Moreover, other data from SNF and SNPs do not seem to support a conclusion of significantly different results. The SNF birds average slightly higher fecundity (0.62 young/pair vs. 0.57 young/pair), but SNPs birds have slightly higher annual survival rates (0.87 vs. 0.80 for males, and 0.84 vs. 0.81 for females). Although none of these differences is significantly different, the general results are consistent with those of Alan Franklin (in press) from his demographic study of the northern spotted owl on the Willow Creek site in the Klamath National Forest. Alan found that birds with access to larger blocks of suitable habitat had slightly lower mortality rates, but those whose home ranges were more patchy (fragmented?) had slightly higher fecundity. Speculation that SNF may be an "ecological sink" for the owls, while SNPs is an "ecological source" (Zabel et al. 1992--cited in the Status Report @ 9) was premature and is not substantiated by data so far available from SNF and SNPs. Over the period from 1990 to 1999, only three individuals--all juveniles--have dispersed between the two study areas, one from SNPs to SNF and two from SNF to SNPs.

B. Second, from the letter to Bradley Powell: "As detailed in the enclosed report, there is strong evidence that the owl's population in the Sierra Nevada is declining precipitously due to habitat loss, degradation, and fragmentation... The same evidence shows that the Forest Service is failing to provide sufficient habitat to ensure viable owl populations..."

The Status Report's link between apparent declines in owl populations and the assertion of habitat loss, degradation, and fragmentation being the cause must be based on the demographic studies and the areas enclosed within their boundaries because evidence of declining populations in the national forests is available only from those areas. Until we have clear evidence that specific logging activities, overall timber harvest, etc. can be linked directly to significantly negative l values and/or to signifcant declines in total counts of owls in the demographic study areas, the true causes of what we are seeing currently cannot be substantiated by science. First, note that items discussed in point 5.A., above, apply here as well. In addition, anecdotal evidence from the SNF demographic study area contradicts a conclusion that the present decline in owl numbers is "...due to habitat loss, degradation, and fragmentation..."

Only one true CASPO sale has been implemented in the SNF study area to date; it was harvested in 1997 and 1998 (after 7-8 consecutive years of l values less than 1 and 3-4 years after l could be confirmed as significantly less than 1); the sale area was on the edge of the study area near the upper elevational limit of owl usage; units treated covered only 303 acres (M. T. Smith, Sierra National Forest silviculturist, pers. comm.), or only 0.2% of the SNF study area; 2.5 million board feet of timber were removed, or about 8,250/acre; and the treated stands remain fully consistent with our understanding of suitable foraging habitat. The only nearby owls did not nest in 1997 or 1998, but they initiated nesting in 1999 and were nearly at the point of fledging nestlings when the violent storm of 2 June dropped 5 inches of snow over their home range. Their nesting attempt failed immediatley after that.

In a study of the possible effects of logging on occupancy and productivity of owl sites in the Big Creek watershed, within the SNF demographic study area, the "Pine Bluff" PAC has been affected on three sides by extensive thinning to create fuelbreaks and begin a long process of establishing small-group selection as a silvicultural prescription in the watershed. The logging occurred in the summer of 1997, with mastication scheduled for 1998 (that has been delayed 1 year), and prescription burning for 1999. The pair produced two young in 1997, began nesting in 1998 but failed when an unusually severe, El Ni–o storm hit in May. The pair has produced two fledglings in 1999 that survived the 2 June storm and are presently still alive.

Finally, the total timber harvest in the Kings River District , which includes the SNF demographic study area, averaged 26.45 million board feet per year from 1990 through 1993, when the number of owls in the study area was on the increase (see Figure 1 @ 5 in the Status Report), but only 8.98 million board feet per year from 1994 through 1998--the period covering the observed decline in numbers of owls in the study area. The difference between these mean volumes is statistically significant (P < 0.001; t-test). Surely this doesn't mean that the owl population increased during the period of higher harvest volumes, but then declined because the harvest volume dropped by 64%, but that is a possible inference (hypothesis) from such data.

Based on information I obtained from the Lassen NF (Philip Sitze, pers. comm., 6-7 July) and Eldorado NF (Wayne Holmstrom, pers. comm., 6 July), the following activities have occurred in their respective demographic study areas since the CASPO Interim Guidlines were implemented: In the Lassen study area--37 timber sales harvested; a total of 14,580 acres treated for an average of 0.7% of the study area's landbase per year (approximating a 143-year rotation); and an average of 3,207 board feet per acre removed (equivalent to about 5 trees/acre @ 24 inches in dbh, but these volumes also included tops, limbs, and trees to 3 inches in dbh). In the Eldorado study area--11 timber sales harvested; a total of 1,392 acres treated for an average of about 0.7% of the study area's National Forest landbase per year (so, again, approximating a 143-year rotation); and an average of 9,357 board feet per acre removed (equivalent to about 14 trees 24 inches in dbh per acre). Without a much better understanding of silviculture, tree-growth projection models, etc., I'm not in a position to judge how well these rates of entry and harvest might affect the long-term maintenance of suitable owl habitat. Certainly provisions would be needed to assure that a baseline density of trees on the order of 250-300 years old would be sustained over the forested landscape. Nor am I able to evaluate the extent to which the volume/acre of timber removed by these sales might require time to become good owl habitat again. This is a question for silviculturists, tree-growth models, etc., and it would depend considerably on the nature of the stands before harvest. The volume per acre on the Eldorado does seem high to me, however, especially considering that the study area is approximately 40% in private ownership and in an identified "Area of Concern" by the CASPO Report @ 47-48.

Second, hazard trees, in my understanding, are typically located along roadways, trails, and in campgrounds. Their removal from those places, therefore, is unlikely to influence fragmentation measurably. Moreover, those locations are not typically where owls roost or select nest trees.

Second, for further clarification, the language implies that all of the numerous projects reviewed involved treatments within PACs and SOHAs. I doubt whether this is true because all such projects would need to invoke the "special circumstances" provision under Adaptive Management Strategies in the CASPO EA (@ III-4). The abundance of such occasions would surely have been widely known, if true. Similar language is found in the body of the text. Although this may be just an oversight in the language here, it is nonetheless inflammatory by implying intentional entry into what are essentially owl reserves.

For my remaining comments on this point, I assume that all, or nearly all, of these projects were implemented in places considered to be probably within owl home ranges but not directly within PACs or SOHAs. The numbers are startling, to say the least, so it seems important to me that we attempt to learn more about this matter. How many of the projects were clearly trivial in their potential impacts? How many were likely to have had some real impact? And how many probably had substantial impact? Without such an analysis, it is not possible to ascertain the full cumulative effects of the multiple projects over the entire Sierra Nevada. I believe this also means that any given BE cannot lightly reach a determination of "...may affect individuals, but is not likely to result in a trend toward Federal listing."

The problem rises to the Regional level, as no single National Forest has a large enough population of spotted owls to assure very-long-term viability. It seems to me that the only way to meet this challenge is to establish a Region-wide database that summarizes the ongoing accumulation of statements to the effect that a project "...may affect individuals, but is not likely to result in a trend toward Federal listing." Projects would need to be ranked according to an evaluation of the likely severity of effects on the individuals that are probably affected. And some Region-wide effort would be needed to identify trigger points beyond which it would no longer be safe to conclude "...but is not likely to result in a trend toward Federal listing."

"Non-select strata" is new usage in my experience. I assume that it refers to "Other Timber Strata," defined in the CASPO report @ 22 as "Stand types used for nesting by the owls, but not significantly selected based on availability..." Eighty-one of the 148 nest sites available in our database at the time the CASPO report was prepared were in Other Timber Strata, and 44 of those were in M3N and M3G strata. The data were not strong enough, however, to establish significant selection of those strata for nesting by the owls in proportion to their availablility over the Sierran landscape. M3N and M3G are highly suitable owl habitat, often for nesting and especially for foraging, and comprise about 25% of all the available, suitable owl habitat in the Sierra Nevada. It would be a serious error to impart a notion that they are unsuitable.

I am puzzled by the apparent contradiction in language of the Status Report, which first seems to criticize the Forest Service for not establishing PACs at newly discovered owl sites and next condemns the agency for adopting "...an outmoded system of reserves...."