Estimates of wear rates in metal bird bands, with applications for survival and movement models of marked individuals

Resumen

Banding allows identification of individual birds, and many investigators use data from recaptured or recovered birds to infer population parameters, including survival rates, migration rates, and recruitment rates. Such analyses often assume that band loss is negligible, but wear rates have been found to differ markedly among different band sizes, band metals, and species in the few cases where wear rates have been determined. However, rapidly wearing bands may be over-reported in the published literature relative to slowly wearing bands for two reasons. First, researchers publish reports of rapidly wearing bands as a warning to other researchers, but may not publish reports of bands wearing slowly. In addition, sampling error may result in over- or underestimation of wear rates. At the extreme, underestimates of wear rates may imply that bands gain mass while worn (i.e., have negative wear rates), whereas extreme overestimates look like rapid wear; only extreme underestimates, not extreme overestimates, are unpublishable due to their a priori implausibility. Hence, the existing literature on band-wear estimates is likely upwardly biased relative to the true distribution of wear rates across species and band metals. Using routinely archived returned bands from the Australian Bird and Bat Banding Scheme for the period from 1963 to 2005, we estimated wear rates of bands applied to 173 species, five band metals, and 236 species/band size/band metal combinations. Band wear rates were generally well-explained by band metal and species functional group, i.e., birds of prey, waterbirds (shorebirds, herons, and ibises), passerines, waterfowl (ducks, geese, and swans), rails, seabirds, parrots, and other non-passerines, but some species had highly divergent wear rates compared to other species in their functional group. We also found that published estimates of wear rates were, on average, more rapid for a given metal type than determined from our analyses. We suggest that publication biases favoring publication of estimates of more rapid wear rates may drive the contrast between estimated wear rates in our analyses and published estimates. El anillamiento permite la identificación de individuos de aves y muchos investigadores usan los datos de aves recapturadas o recuperadas para inferir parámetros poblacionales incluyendo tasas de supervivencia, tasas de migración y tasas de reclutamiento. Estos análisis asumen que la perdida de los anillos puede ser ignorada, pero en los pocos casos en los que las tasas de desgaste ha sido determinada, se ha encontrado que las tasas de desgaste difieren marcadamente entre anillos de diferentes tamaños, el tipo de metal y las especies en las que son usados. Sin embargo, el rápido desgaste de los anillos puede ser sobre reportado en la literatura en relación con los anillos que se desgastan más lentamente. Primero, los investigadores publican el reporte del desgaste rápido de los anillos como advertencia para otros investigadores, pero pueden no publicar los reportes de anillos que se desgastan lentamente. Adicionalmente, el error de muestreo puede resultar en sobre o subestimación de las tasas de desgaste. En el extremo, la subestimación de las tasas de desgaste puede implicar que los anillos incrementan en masa mientras sean usados (i.e., tienen tasas de desgaste negativo), mientras que sobreestimaciones extremas puede observarse como desgaste rápido; los resultados son no publicables solo en el extremo de subestimación, no en el extremo de sobreestimación, debido a su falta de plausibilidad a priori. Por esta razón, la literatura existente en estimados de desgaste de anillos probablemente esta sesgada hacia arriba con respecto al a distribución real de las tasas de desgaste a través de especies y anillos metálicos. Usando los archivos de anillos que rutinariamente son devueltos al “Esquema de Anillamiento de Aves y Murciélagos de Australia” para el periodo comprendido entre 1963 y 2005, estimamos las tasas de desgaste de anillos utilizados en 173 especies, cinco tipos de metal y 236 combinaciones de especies/tamaño/tipo de metal. Las tasas de desgaste fueron generalmente bien explicadas por el tipo de metal y el grupo funcional de la especie, i.e. rapaces, aves acuáticas (aves playeras, garzas e ibis), paseriformes, aves acuáticas de cacería (patos, gansos y cisnes), tinguas, aves marinas, loros y otras aves no paseriformes, pero algunas especies tuvieron tasas de desgaste altamente divergentes comparadas con otras especies en su grupo funcional. También encontramos que los estimados publicados de las tasas de desgaste fueron, en promedio, mas rápidos para un metal determinado que lo estimado por nuestros análisis. Sugerimos que los sesgos de publicación favoreciendo la publicación de los estimados de tasas de desgaste mas rápida puede ser responsable por el contraste entre las tasas de desgaste estimadas por nuestros análisis y los estimados publicados. Table S1. Starting masses of each band type as used in this study, inferred from samples of unworn retained and returned bands from the ABBBS archives. Table S2. Metallurgical details for alloy and incoloy bands. Table S3. Estimated wear rates and band wear parameters for all band type/species measured, including those with N < 3. Fig. S1. Expected bias resulting from under-sampling of rapidly wearing bands. Fig. S2. Monel bands can show differing apparent wear rates within species, apparently driven by differences in the bands. Fig. S3. Stainless steel bands appear to show heterogeneous wear rates by functional group, with bands worn by passerines losing a greater proportion of their mass per unit time than bands worn by other groups. Fig. S4. Aluminum bands appear to wear in two clusters of rates by functional group. Aluminum bands on seabirds, passerines, and ducks, geese, and swans wear at a fairly rapid rate (~3%/yr), whereas bands on birds of prey, waders, herons, and ibises, and other non-passerines wear at a considerably slower rate (~0.5%/yr). Fig. S5. Alloy bands on passerines appear to wear at a fairly rapid rate (~4%/yr). Fig. S6. Stainless steel bands on seabirds uniformly show little wear unless they are worn by Kelp Gulls (Larus dominicanus) or, possibly to a lesser extent, by Pacific Gulls (L. pacificus). Fig. S7. Stainless steel bands on waders, herons, and ibises show a split between the wear rate for Australian White Ibises (Threskiornis molucca) and Pied Oystercatchers (Haematopus longirostris); other species are data-poor in comparison, but Royal Spoonbills (Platalea regia) appear to have a particularly low wear rate. Fig. S8. Stainless steel bands on parrots show highly variable apparent wear on Galahs (Eolophus roseicapillus); other species have small sample sizes or limited durations on birds.In addition to the above Supplementary Tables and Figures, the following are available by request from the corresponding author, or on Github (https://github.com/SMBaylis/BandWear_JFO): (1) R code for Band Wear Table v0.4.R, (2) Band Wear Table v0.4.R, (3) Literature comparisons, (4) Taxonomic orders, (5) Worn band masses, (6) Guilds by species, (7) Band details, (8) Unworn bands, (9) Full worn bands, (10) Unworn band masses, and (11) Unworn bands. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.