Hoǁ ŵanLJ saŵples needs an edžhausiǀe ďird inǀentorLJ in forest enǀironŵents? Rosario Balestrieri ϭ* , Marco Basile ϭ , Tiziana Altea Ϯ , Giorgio Matteucci ϭ , Mario Posillico ϭ,Ϯ ϭ Isituto di Biologia Agroaŵďientale e Forestale del CNR Via Salaria kŵ Ϯ9,ϯϬϬ, ϬϬϬϭϱ Monterotondo SĐalo RM Ϯ Corpo Forestale dello Stato, UiĐio Territoriale Biodiǀersità di Castel di Sangro, Via Sangro, ϰϱ-ϲϳϬϯϭ Castel di Sangro ;AQͿ *rosario.ďalestrieri@iďaf.Đnr.it INTRODUCTION )nvestigating the relationship between birds and forests is a key argument for sustainable forestry and biodiversity conser- vation, as birds are often the most numerous group of forest vertebrates, both as number of species and individuals ȋDeGraaf et al. ͳͻͻ͸Ȍ. Changes in forest characteristics ȋe.g., tree age or canopy coverȌ, due to forestry practices, could af- fect the composition of bird communities ȋCaprio et al. ʹͲͲͺ, Gil-Tena et al. ʹͲͲͺ, White et al. ʹͲͳ͵Ȍ. On the other hand, the composition of bird communities could be a powerful tool in predicting how forestry practices affect biodiversity. An ex- haustive bird inventory is the starting point of a reliable analysis of such community. The actual meaning of Dzexhaustivedz depends on the proposed research objective, but hardly ever an exhaustive sampling can be performed with a negligible cost in terms of both time and money. )t usually requires a high number of replicated sampling points, proportionally to the size of the study area ȋDickson ͳͻ͹ͺ, Toms et al. ʹͲͲ͸Ȍ. Such a sampling design could be sometimes difficult to afford be- cause wildlife monitoring is often bounded by limited financial resources. Skilled personnel is costly and, as it is essential for an unbiased data collection, project manager often has to minimize the sampling effort. We carried out our study within the project L)FE ManFor CBD ȋwww.manfor.euȌ, which aims to test and verify in the field the effectiveness of forest management options in meeting multiple objectives ȋtimber production, environment protection and biodiversity conservation, etc.Ȍ, providing data, guidance and indications of best -practice. Therefore, we search for an optimal sampling scheme to describe interaction between forest bird communities and forest ecosystems. )ndeed, our aim was to seek the optimal compromise between the magnitude of sampling effort and a proxy of an exhaustive bird inventory. METHODS Data were collected in ͷ forest stands in ͷ ar- eas from northern to southern )taly, each one extended for about ͵Ͳ ha: Norway spruce ȋPicea abiesȌ forest in Julian Alps ȋ~͵Ͳ haȌ; Norway spruce – silver fir ȋAbies albaȌ forest in Dolomites ȋ~ʹͷ haȌ; and ͵ beech ȋFagus sylvaticaȌ forests in Venetian Prealps ȋ~͵͵ haȌ, central Apennines ȋ~͵Ͳ haȌ and south- ern Apennines ȋ~͵Ͳ haȌ. We randomly select- ed ͳͻ – ʹͶ points in each area, stratified ac- cording to the different silvicultural options to be applied, ͳͳͳ overall ȋtab. ͳȌ. Every point was at minimum distance of ͳʹͲ m from the others, in order to minimize spatial autocor- relation. Every area was sampled ͷ times, but the one in southern Apennines which was sampled Ͷ times. Point count was performed through aural and visual survey and lasted ͷ minutes, during which species and individual numbers were recorded. To obtain compara- ble data we sampled over a short time frame during the peak of bird activity ȋMay-JuneȌ, from dawn to ͳͳ:ͲͲ. The capability of the sampling design in detecting species was compared with the sample coverage estima- tor developed by Chao et al. ȋʹͲͳʹȌ and im- plemented in the online software iNext ȋ(sieh et al. ʹͲͳʹȌ which we also used to car- ry out rarefaction and extrapolation curve, based on ͶͲ knots, and furnished with a ͻͷ% confidence interval of ͳͲͲ replications of bootstrap resampling ȋChao et al. ʹͲͳ͵Ȍ. Fi- nally, we calculated for every area the median value of the percentage of species detected at every sampling occasion and ran out a non- parametric Bonferroni corrected Mann- Whitney test to search for significant differ- ence in detectability among areas. RESULTS Sample coverage estimators was always > Ͳ.ͻͷ, showing that an exhaustive sur- vey has been performed in every area after Ͷ/ͷ sampling visits ȋtab. ʹȌ. Rarefac- tion and extrapolation curves suggest that in every area observed species rich- ness has reached the plateau of the curve ȋfig. ͳȌ. Since the sample coverage was high, we assume that the cumulative percentage of species detected in the last occasion can be considered as ͳͲͲ% of current species. Moreover, Bonferroni corrected Mann-Whitney test showed that during every sampling occasion in each area the percentage of species detected did not differ significantly ȋLS < Ͳ.ͲͷȌ. The mean cumulative percentage of species detected from first to last oc- casion was: ͹͵.͸Ͷ %, ͺ͹.͹ʹ %, ͻʹ.ͻ͸ %, ͻͷ.͹ͺ %, ͳͲͲ% respectively ȋfig. ʹȌ. Ta- ble ͵ lists the percentage of species detected at every sampling occasion for each area, reaching the ͻͲ% threshold after ͵ sampling occasions. Figure ͳ: rarefaction and extrapolation curves, divided by forest, performed by the online software iNEXT ȋ(sieh et al. ʹͲͳʹȌ, based on ͶͲ knots and furnished with a ͻͷ% confidence interval of ͳͲͲ repli- cations of bootstrap resampling ȋChao et al. ʹͲͳ͵Ȍ. First graph of each series shows sample coverage trend in relation to new species added to the area inventory. Second graph of each series shows how many samples are needed for a full sample coverage. Third graph of each series shows species richness at every sampling occasion. Continuous line = realized curve; dashed line = predicted curve; shaded area = ͻͷ% C). Ringraziaŵeni: si ringraziano le UTB di Tarǀisio, Vitorio Veneto, Castel di Sangro, Mongiana e il Coŵune di Lorenzago di Cadore. Photo: Daǀide De Rosa Table 1: Study areas description. Name Extension (ha) Forest type N. of points Pian Parrocchia Campo di Mezzo Nat. Reserve, Pian di Cansiglio (TV) ~ 33 Beech, high growth (Fagus sylvatica) 21 Chiarano – Sparvera regional forest (AQ) ~ 30 Beech, high growth 23 Valdescura, Lorenzago di Cadore (BL) ~ 25 Mixed spruce (Picea abies) and silver fir (Abies alba) 19 Marchesale Nature Reserve, Mongiana (VV) ~ 30 Beech, high growth 24 Tarvisio, (UD) ~ 30 Mixed conifer forest – 54% spruce 24 Table 2: Sample coverage estimator (Chao et al. 2012) for every area, number of sampling occasion and number of detected species. Area Sampling occasion Detected species E. sample coverage Cansiglio 5 27 0.9824 Chiarano 5 26 0.976 Lorenzago 5 28 0.9587 Mongiana 4 21 0.994 Tarvisio 5 30 0.9641 Table 3: Percentage of detected species in each area during each sampling occasion. There were only 4 sampling occasion in Mongiana. Area Sampling occasion 1 2 3 4 5 Cansiglio 60.7 78.6 89.3 89.3 100 Chiarano 80.8 96.2 96.2 100 100 Lorenzago 64.3 78.6 89.3 92.9 100 Mongiana 85.7 95.2 100 100 X Tarvisio 76.7 90 90 96.7 100 LITERATURE Caprio E, Ellena I, Rolando A (2008). Assessing habitat/landscape predictors of bird diversity in managed deciduous forests: a seasonal and guild-based approach. Biodivers. Conserv. 18: 1287 – 1303. Chao A, Gotelli NJ, Hsieh TC, Sander EL, Ma KH, Colwell RK, Ellison AM (2013). Rarefaction and extrapolation with Hill numbers: a framework for sampling and estimation in species diversity studies . Ecological Monographs: http://dx.doi.org/10.1890/13-0133.1 Chao A, Jost L (2012). Coverage-based rarefaction and extrapolation: standardizing samples by completeness rather than size. Ecology 93: 2533 – 2547. DeGraaf R, Miller RI (1996). Conservation of faunal diversity in forested landscapes. Chapman & Hall, London. Dickson JG (1978). Comparison of breeding bird census techniques. American Birds 32 (1): 10 – 13. Gil-Tena A, Torras O, Saura S (2008). Relationships between forest landscape structure and avian species richness in NE Spain. Ardeola 55 (1): 27 – 40. Hsieh TC, Ma KH, and Chao A (2013). iNEXT online: interpolation and extrapolation (Version 1.0) [Software]. Available from http://chao.stat.nthu.edu.tw/blog/software-download/. Toms JD, Schmiegelow FKA, Hannon SJ, Villard M-A (2006). Are point counts of boreal songbirds reliable proxies for more intensive abundance estimators? The Auk 123 (2): 438 – 454. White AM, Zipkin EF, Manley PN, Schlesinger MD (2013). Conservation of avian diversity in the Sierra Nevada: moving beyond a single-species management focus. PLoS ONE 8(5): e63088. doi:10.1371/journal.pone.0063088. CANSIGLIO CHIARANO LOREN)AGO MONGIANA TARVISIO Figure ʹ: cumula- tive mean percent- age of additional species detected af- ter every sampling occasion, among each area. Any re- markable improve- ments are found af- ter the third occa- sion. DISCUSSION Therefore, we consider that ͵ sampling occa- sions could characterize in a satisfactory way the bird community within the studied forest types in the pre-treatment period. Our study provides an useful methodological tool for planning bird inventories in forest environ- ments when personnel and financial resource are limited, leading to a thoughtful fund man- agement IX Congresso Nazionale SISEF—Bolzano ϮϬϭϯ