CTFS-ForestGEO ARTHROPOD INITIATIVE 2008-2015: Monitoring barcodes? Long-term monitoring of tropical arthropods Yves Basset, Smithsonian Tropical Research Institute, [email protected] Currently: 61 sites in 24 countries 6 million of trees monitored, representing 10,000 species 10 science initiatives: Arthropod monitoring: 9 sites CTFS-ForestGEO ARTHROPOD INITIATIVE Aims: to monitor key insect assemblages over the long-term at CTFS sites and to study insect-plant interactions across the CTFS network At each CTFS site, 3 phases: - baseline survey to identify common species - monitoring (modeled on baseline survey) - interaction studies (different set of protocols) Focus on a priority set of assemblages chosen for their • ecological relevance • taxonomic tractability • ease of sampling Monitoring Interactions only Priority assemblages Litter ants: key organisms in tropical forests and often key predators [Formicidae] Selected moths and butterflies: caterpillars leaf-chewers, adults often pollinators [Rhopalocera, Geometridae, Arctiinae & Pyraloidea] Bees: important pollinators of many tropical trees [Apidae Euglossini] Termites: important decomposers in tropical forests [Isoptera] Tephritid fruit-flies: seed (fruit) predators [Tephritidae] Seed predators: important influence on fruit/seed survival (whole guild) [Varia] Full suite of 15 taxa studied at BCI, Panama Methods: baseline survey & monitoring Litter ants: extraction from litter with Winkler Bees: attraction to chemical baits, (only Neotropical sites) Tephritid fruit-flies: baited McPhail traps (not in the Neotropics) Moths and other taxa: light traps Termites: light traps & hand search in quadrats Butterflies: walking transects INTERACTION STUDIES • KHC (2010): Effects of litter composition on ants • BCI (2010), KHC (2013), WAN (2013): Insect seed predation: quantitative food webs TAXONOMY • Local reference collections often managed by “parataxonomists” • DNA barcoding • Collaborating experts (in-country or abroad) But taxonomic impediment high in the tropics! Example: BARRO COLORADO ISLAND PANAMA (BCI) Progress in DNA barcoding since 2009 STATUS 2008: baseline survey 2009-2015: on-going monitoring (7th year) Collections: 36,388 pinned specimens; 1,864 spp. 68% of spp. with Barcode Index Numbers (BINs) Four full-time research assistants, based at STRI MAIN USE OF DNA BARCODING • Used to refine delineation of insect morphospecies and identifications of common, abundant species, likely to be tractable in the long-term • Currently > 10,000 sequences: Panama (7,414), Thailand (1,053), Papua New Guinea (1,567) • Some of these areas have extensive barcode librairies (PNG 25,000 barcodes for Lepidoptera), others not (Thailand): approach for barcoding needs to be different for each site INTEREST OF DNA BARCODING • Refinement of reference collections BCI: Arctiinae, Geometridae, etc: best collections in Panama • Baseline and ecological surveys (Guanacaste: Janzen et al.) Of interest when monitoring protocols are similar to the most efficient collecting methods for focal taxa. For example: Saturniidae collected at light on BCI: 40 (a) 30 20 10 Mean cumulative no. of species of no. cumulative Mean 0 0 200 400 600 800 Individuals Species accumulation curve OK Documentation of local fauna Representative tree INTEREST OF DNA BARCODING DNA barcoding allows: • Matching sexes in case of sexual dimorphism • Identification of caterpillars, pupae and reared parasitoids • Matching casts of social insects (sometimes across different protocols) Cornitotermes walkeri BOLD:AAP9751 Alate, Soldier, light trap termite transect INTEREST OF DNA BARCODING DNA barcoding allows monitoring functional groups important to the forest ecosystem but not well known taxonomically Flatidae (sap-sucking bugs): 10% of BOLD sequences from BCI 300 FLATIDAE, BCI 250 raw data 200 2009 2015 150 100 No. of individuals No. of 50 0 2 3 4 5 8 9 10 11 14 15 16 17 20 21 22 23 26 27 28 29 32 33 34 35 38 39 40 41 Light trap survey RESULTS: SELECTED EXAMPLES • Yearly results: annual indices (butterflies & ants) • Population dynamics (saturniid moths and butterflies) • Yearly changes in assemblages (several taxa) Immediate significance vs. Interpretation of long chronosequences Annual indices, BCI, Year 2011 50 All butterflies: 197 ± 9.8, 5.0% 7.1 precision, % 45 40 60,000 insects collected: 35 10.7 17,000 focal individuals (910 spp.) 30 25 19.2 14.3 20 For 56 spp. we can estimate annual indices 15 with good precision 17.2 10 17.9 15.7 5 56 spp. = 6% of total spp. but 0 Pierella luna Itaballia Detritivora Itaballia Cithaerias Mesosemia Phanus Mean Mean ind. and per site s.e. censued (annual index) 55% of total abundance of focal taxa pandosia hermodora demophile pireta hesperina marshalli 0.9 All ants: 1.00 Annual indices: 0.8 Non-social insects: mean per site (n=10) 0.7 Precision = s.e./mean 0.6 0.5 (< 20% very good, economic entomology) 0.4 0.3 Social insects: occurrence in samples, Occurence and 95% c.l. Occurence 0.2 transects or quadrats 0.1 Precision = 95% c.l. on occurrence data, 0 assuming a binomial distribution 3 1 rugiceps Pheidole ruidum Pheidole Pyramica gundlachi multispina minutus Ectatomma Hypoponera Wasmannia distinguenda auropunctata Cyphomyrmex Solenopsis sp. Solenopsis sp. Pyramica zeteki Annual indices Immediate significance: • Common spp. can be monitored with relative precision, even in tropical rainforests • Few long-term monitoring programs in the tropics (butterflies: 10-11 years: Leidner et al. 2010, Grøtan et al. 2012) • Indices for social insects need to be reported differently than for non-social insects (refinements needed for social insects; geometric mean for non-social insects) • BCI: all common spp. (with annual indices) have BINs Population dynamics, BCI, 2009-2014, saturniid moths Rhescyntis 2011 Hylesia sp. 2YB 2014 20 2 20 4.5 All Saturniidae 2011 hippodamia (LT) Time-series models 200 18 18 n.s. (ET) 2013 4 180 All saturniids 2012 16 (TRIM): 16 3.5 160 1.5 14 14 2009 3 (1) No time effects 140 2013 12 12 2010 2.5 120 2009 2010 (2) Linear trend 10 1 10 2011 100 2 8 2009 2012 2013 8 (3) Effect for each time 80 2012 1.5 6 2010 6 60 2014 1 point 0.5 4 4 40 2014 2 0.5 2 (separate parameters 20 0 0 0 0 0 0 10 20 30 40 50 60 70 for each time-point) 0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70 Periphoba 2011 Automeris 2011 25 5 30 Automeris 2011 sp. 1YB fieldi 20 3 zuganaDHJ03 (ET) 4.5 n.s. 18 (ET) 25 20 4 2.5 16 3.5 20 14 2009 2 15 3 12 2010 2012 2.5 15 10 1.5 10 2 2010 8 2012 2009 10 2009 1 1.5 2012 2013 2014 6 2010 2013 2014 2013 5 1 5 4 2014 0.5 0.5 2 0 0 0 0 0 0 10 20 30 40 50 60 70 collected per survey per collected 0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70 Pseudodirphia 2011 18 7 Automeris eupanamensis Pseudoautomeris 10 1.5 30 2011 6.5 salmonea vanschaycki 16 (ET) 6 n.s. 9 (NT) 14 5.5 25 8 5 2009 2010 2011 2012 2013 2014 12 7 4.5 index Annual 20 1 6 10 4 2012 3.5 15 5 8 2014 3 4 2.5 6 10 0.5 2009 2010 2 2010 3 4 2012 2013 1.5 2 2013 5 1 2 2009 2014 1 0.5 0 0 0 0 0 0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70 Automeris 2011 Oxytenis 14 6 Oxytenis 7 2013 6.5 belti 8 1.5 sp. 1YB naemia 6 (ET) 5.5 (ET) 12 (NT) 6 5 7 5.5 4.5 5 10 6 2009 2010 2011 2012 2013 2014 5 Number of individuals individuals of Number 4 2010 4.5 2014 1 4 Model 8 3.5 5 4 3.5 3 2014 4 3 6 2010 2012 2.5 3 2.5 2 3 2012 4 2009 0.5 2 2 1.5 2013 2 1.5 Raw data 1 2009 2011 2 1 1 1 0.5 0.5 0 0 0 0 0 0 0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70 Months since January 2009 Population dynamics, BCI, 2009-2014, butterflies Exclusive feeders on: 180 3 Herbs (49 spp.): each time point 2012 160 Trees (35 spp.): model n.s. Liana (32 spp.): model n.s. 2011 2013 2.5 140 120 2 2010 2014 100 1.5 80 2009 60 1 indice Annual Number of individuals of Number 40 0.5 20 0 0 0 10 20 30 40 50 60 70 Months since January 2009 Different feeding guilds, moderate increase for herb feeders since 2009 (p<0.01) Population dynamics Immediate significance: • We can detect significant (short-term) trends • Nearly a quarter of species show significant changes with time (different groups tested) • Population dynamics may depend on the ecology of species • Species identity is crucial, therefore BINs are essential Changes in assemblages, BCI, 2009-2014 Butterflies 0.8 2014 Geometridae 1 2009 283 species 2013 254 species 0.6 0.8 Transects Light traps 0.6 0.4 2010 Matrices Spp. x Years 2011 0.4 0.2 NDMS 0.2 0 0 Different protocols, different taxa -2 -1.5 -1 -0.5 0 0.5 1 1.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 -0.2 2009 -0.2 2013 & 2014 6 out of 16 assemblages 2010 -0.4 -0.4 2012 -0.6 (different guilds) -0.6 2011 -0.8 2012 Stress= 0.00 Stress= 0.00046 -0.8 -1 Euglossine bees 1 Flatidae 0.8 Mantel tests, p-values: 24 species 0.8 2009 & 2011 2010 27 species 0.6 Baits 0.6 Light traps 0.4 2013 Geometridae Bees Flatidae Ants Termites 0.4 0.2 2013 & 2014 Butterflies 0.152 0.482 0.815 0.044 0.617 0.2 2014 0 -2 -1.5 -1 -0.5 0 0.5 Geometridae1 *** 0.240 0.593 0.432 0.043 -0.2 0 2012 -1.5 -1 -0.5 0 0.5 1 1.5 2 Bees *** 0.094 0.987 0.221 -0.4 -0.2 2009 Flatidae *** 0.989 0.210 -0.6 2011 -0.4 -0.8 Ants *** 0.849 -0.6 -1 -0.8 2012 -1.2 2010 Stress= 0.035 -1 Stress= 0.00040 -1.4 2012 0.8 Termites 1.5 16 species 0.6 2014 Quadrats 2012 2011 1 0.4 Trajectories are mostly independent 2011 0.2 0.5 0 2010 -1.5 -1 -0.5 0 0.5 1 1.5 -0.2 2010 Directional changes are few 0 -0.4 2013 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 -0.6 2013 -0.5 -0.8 2014 -1 Formicidae 2009 -1 2009 -1.2 124 species Stress= 0.00049 Stress= 0.074 -1.4 Winkler -1.5 Changes in assemblages, BCI, 2009-2014 Geometridae 1 2009 Flatidae 0.8 254 species 0.8 27 species 2009 & 2011 Light traps 0.6 Light traps 0.6 0.4 2010 Matrices Spp.