The URL can be used to link to this page

Home My WebLink AboutUSGS_Macroinvertebrate_AssemblageMacroinvertebrate Assemblage Correspondence to Hydrologic Classes T.F. Cuffney, Ph.D. U.S. Geological Survey North Carolina Water Science Center Analysis issues •Uneven distribution among classes •Rare taxa •Multiple seasons •Ordinal data •Ambiguous taxa •Lowest taxonomic level Uneven distribution among classes Data sets used in analyses: 1. All classes and sites 2. All classes except E a. All sites b. 10 random sites: A, B, D Rare taxa 0 20 40 60 80 100 120 140 160 123456789101112131415161718192021 No . of ta x a Occurrence (no. sites) 22 % 13 % 114 55 % Create 5 data sets all taxa taxa at > 5 sites > 10 sites > 15 sites > 20 sites Multiple seasons 0 5 10 15 20 25 30 35 40 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Si t e s 56 % Ordinal Data and Ambiguous Parents •Ordinal data: Absent < Rare < Common < Abundant 0 13 10 Ambiguous parents: Site 1Site 2Site 3Site 4 Order Ephemeroptera 1 Family Baetidae 1 10 Genus Baetis 33 1 Species B. flavistriga 133 B. pluto 10 3 Resolving ambiguous taxa and recoding data Site 1 Site 2 Site 3 Site 4 Order Ephemeroptera 1 Family Baetidae 1 10 Genus Baetis 33 1 Species B. flavistriga 133 B. pluto 10 3 Site 1 Site 2 Site 3 Site 4 Order Ephemeroptera Family Baetidae Genus Baetis 4 Species B. flavistriga 1.3 8 4 B. pluto 12.7 8 1. Remove ambiguous parents ≥ Order 2. DPAC‐s: Distribute Parent Among Children by sample Site 1 Site 2 Site 3 Site 4 Order Ephemeroptera Family Baetidae Genus Baetis 3 Species B. flavistriga 133 B. pluto 10 3 Site 1 Site 2 Site 3 Site 4 Order Ephemeroptera Family Baetidae Genus Baetis 1 Species B. flavistriga 111 B. pluto 11 Qualitative (P/A) recoding Ordinal recoding Lowest taxonomic level Ambiguous taxa Lowest Richness Abundance Total taxa level No. % No. % taxa Species 116 19.0 5,819 17.6 610 Genus 27 6.6 712 2.6 408 Analysis with lowest taxa level: Species Genus Data preparation steps 1. Set lowest taxa level: Genus, Species 2. Remove rare taxa: 0, 5, 10, 15, 20 sites 3. Remove ambiguous parents ≥ Order 4. Resolve ambiguous taxa: Distribute abundance of ambiguous parents among children, proportionately Resolve ambiguities separately for each sample 5. Ordinal and Qualitative (P/A) data sets 6. 10 site simulations for clusters A, B, and D Analysis •ANOSIM: tests correspondence between hydrologic classes and invertebrate assemblages. •Indicator Value Analysis: identifies taxa that differentiate among hydrologic classes. ANOSIM •ANOSIM: analysis of similarity, Primer 6, Clarke & Gorley 2006 •1‐way ANOVA analogue based on non‐parametric permutation procedures •Hydrologic classes = treatments •Analysis is based on resemblance (similarity or dissimilarity) of assemblages between sites •Rank similarities between samples in the underlying triangular similarity matrix Similarity based on assemblages Site 1 Site 2 Site 3 Site 4 Site 5 Spp 1 000211 Spp 2 70 0 47 0 72 Spp 3 00000 Spp 4 55 99 0 93 0 Spp 5 011490 0 Spp 6 0 0 76 73 3 Spp 6 60 0 60 0 43 Spp 7 32 0 0 0 36 Spp 8 09748048 Data matrix Site‐by‐species Site 1 Site 2 Site 3 Site 4 Site 5 Site 1 Site 2 29.20 Site 3 44.31 40.10 Site 4 31.15 42.03 28.62 Site 5 65.32 25.01 65.36 16.91 Resemblance matrix Similarity Quantitative: Kendal rank correlation Qualitative (P/A): Sorenson index Examines averages within & between classes of the resemblance matrix Site 1Site 2Site 3Site 4Site 5Site 6Site 7Site 8Site 9 SiteClassAAABBBCCC Site 1A ‐‐‐ Site 2A 33‐‐‐ Site 3A 8 7 ‐‐‐ Site 4 B 22 11 19 ‐‐‐ Site 5 B 66 30 58 65 ‐‐‐ Site 6B 44 3152829‐‐‐ Site 7 C 23 16 5 38 57 6 ‐‐‐ Site 8 C 9 34 4 32 61 10 1 ‐‐‐ Site 9 C 48 17 42 56 37 55 51 62 ‐‐‐ Within class Test procedure •Test statistic: Where: = average rank similarity between classes = average rank similarity within class M = n(n‐1)/2, n= total samples •Ho: no differences in the composition of assemblages among classes (R=0) •Permutations: random relabeling samples Rperm •Significance: % of Rperm > R Global and Pair‐wise Tests •Global test of significance: a significant result means there are differences somewhere that should be examine further. •Pair‐wise test of significance –Extract pairs of classes –Re‐ranked –Repeat test ANOSIM results: Global DPAC‐S remove ambig. parents ≥ Order Species Genus All sites and classes ns ns Without class E All sites in A, B, and D Rare taxa ≥ 0nsns ≥ 5nsns ≥ 10 ns ns ≥ 15 ns ns ≥ 20 ns ns Random 10 sites in A, B, and D Rare taxa ≥ 0nsns ≥ 5nsns ≥ 10 ns ns ≥ 15 ns ns ≥ 20 ns ns Quantitative data sets ANOSIM results: Global DPAC‐S remove ambig. parents ≥ Order Species Genus All classes 0.22 0.21 W/o class E All sites in A, B, and D Rare taxa ≥ 0 0.22 0.26 ≥ 5 0.23 0.27 ≥ 10 0.22 0.46 ≥ 15 0.21 0.25 ≥ 20 0.25 0.33 Random 10 sites in A, B, and D Rare taxa ≥ 0 0.28 0.25 ≥ 5 0.29 0.37 ≥ 10 0.20 0.42 ≥ 15 0.27 0.32 ≥ 20 0.34 0.36 Qualitative (P/A) data sets Pair‐wise tests: Genus, P/A, ≥ 10 ABCDFG A ‐‐‐‐‐ B 2.8 ‐‐‐‐‐ C 0.2 3.6 ‐‐‐‐‐ D 12.5 100 14.3 ‐‐‐‐‐ F 0.1 1.8 16.5 10 ‐‐‐‐‐ G 0.8 78.6 0.2 57.1 0.02 ‐‐‐‐‐ P‐values: p ≤ 5 % shown in yellow ANOSIM interpretation •Quantitative: correspondence with hydrologic classes is not significantly different from a random assignment of sites to classes. •Qualitative (P/A): correspondence with hydrologic classes significantly better than a random assignment of sites to classes. However, correspondence is relatively low. Unanswered question Is there a hydrologic classification that optimizes the correspondence to the distribution of assemblages? Fidelity analysis Identify taxa strongly associated with a priori groups (hydrologic clusters) Fidelity analysis methods •Indicator value analysis (Dufrene and Legendre 1997): quantitative data •Fidelity analysis (Tichy and Chytry 2006): qualitative (P/A) data Indicator Value index j = cluster abundance of species i in cluster j ‐‐specificity Bij is occurrence of species i in cluster j – fidelity : Fidelity Where: N = number of sites in the data set Np = number of sites in target site group n = number of occurrences of taxon in the data set np = number of occurrences of taxon in the target site groups Fidelity analysis data sets DPAC‐S, remove ambiguous parents ≥ Order, no rare taxa removed: –Lowest taxa level = Species •Quantitative: 0, 1, 3, 10 •Qualitative (P/A): 0, 1 –Lowest taxa level = Genus •Quantitative: 0, 1, 3, 10 •Qualitative (P/A): 0, 1 Global results •Species: –Quantitative: p < 0.001 –Qualitative: p < 0.001 •Genus: –Quantitative: p < 0.001 –Qualitative: p < 0.001 Hydro Class A Species Genus Taxon Quant Qual Quant Qual Amnicola X Amnicola X Isochaetides freyi X Isochaetides X Palaemonetes paludosus XX Palaemonetes XX Gammarus fasciatus XX Gammarus X Acerpenna pygmaea XX Acerpenna XX Neurocordulia obsoleta XX Neurocordulia XX Epicordulia princeps X Epicordulia X Erythemis XX Enallagma XX Enallagma XX Pelocoris X Pelocoris X Belostoma X Belostoma X Phylocentropus XX Phylocentropus XX Macrostemum X Macrostemum X Nectopsyche exquisita X Nectopsyche X Oecetis nocturna X Molanna X Molanna XX Coptotomus X Coptotomus X Laccophilus XX Laccophilus XX Peltodytes XX Peltodytes XX Enochrus XX Enochrus XX Tribelos X Labrundinia pilosella xX Labrundinia X Hydro Class B Species Genus Species Genus Taxon Quant Qual Taxon Quant Qual Taxon Quant Qual Taxon Quant Qual Serratella deficiens X Serratella X Pteronarcys XX Serratella serratoides X Ceratopsyche bronta X Ceratopsyche X Paraleptophlebia X Paraleptophlebia X Ceratopsyche sparna X Acentrella X Dolophilodes X Dolophilodes X Baetis pluto X Goera X Goera X Epeorus vitreus XXEpeorus XX Pycnopsyche XXPycnopsyche XX Heptagenia marginalis X Neophylax XXNeophylax XX Stenacron pallidum X Glossosoma X Glossosoma X Stenonema ithaca X Rhyacophila fuscula XX Stenonema pudicum XX Optioservus ovalis X Lanthus X Lanthus XX Microtendipes x Leuctra XXLeuctra XX Diamesa X Diamesa X Tallaperla XXTallaperla XX Parametriocnemus X Parametriocnemus X Acroneuria abnormis X Antocha X Antocha X Paragnetina immarginata X Paragnetina X Dicranota XXDicranota XX Isoperla holochlora X Atherix X Malirekus hastatus X Malirekus X Hydro Class C Species Genus Taxon Quant Qual Taxon Quant Qual Dugesia tigrina XXDugesia XX Elliptio complanata X Elliptio X Ferrissia X Ferrissia X Limnodrilus XXLimnodrilus X Tricorythodes XXTricorythodes XX Baetis intercalaris X Baetis X Heptagenia XX Stenonema exiguum XX Stenonema integrum X Stenonema ithaca X Gomphus X Gomphus X Macromia XXMacromia XX Paragnetina X Pteronarcys dorsata X Corydalus cornutus XXCorydalus XX Macronychus glabratus X Macronychus X Hydropsyche X Hydropsyche X Brachycentrus numerosus X Paracladopelma XXParacladopelma XX Polypedilum X Polypedilum X Robackia demeijerei XXRobackia X Cricotopus bicinctus XXCricotopus XX Lopescladius X Lopescladius X Thienemanniella X Nilotanypus X Ablabesmyia XX Hydro Class D Species Genus Taxon Quant Qual Taxon Quant Qual Progomphus obscurus X Progomphus X Hydro Class E •Eliminated, only 1 site in class. Hydro Class F Species Genus Taxon Quant Qual Taxon Quant Qual Serratella serratoides X Serratella X Baetis pluto X Heterocloeon XXHeterocloeon XX Hetaerina X Argia X Boyeria vinosa X Boyeria XX Hagenius brevistylus X Hagenius X Hetaerina XX Perlesta X Perlesta X Agnetina X Ranatra X Ceratopsyche morosa XXCeratopsyche X Ceratopsyche sparna X Hydropsyche phalerata X Hydropsyche venularis XX Psychomyia nomada X Oecetis avara X Promoresia elegans XXPromoresia XX Psephenus herricki X Psephenus X Anchytarsus X Cryptochironomus XXCryptochironomus X Microtendipes rydalensis X Paratendipes albimanus X Brillia sera XX Cardiocladius obscurus XXCardiocladius XX Cricotopus (Cricotopus)X Antocha X Antocha X Hydro Class G Species Genus Taxon Quant Qual Taxon Quant Qual Lirceus XXLirceus XX Prostoia X Prostoia X Ironoquia XXIronoquia XX Ptilostomis X Ptilostomis X Parachironomus X Parachironomus X Fidelity analysis: conclusions •Can extract taxa that are indicative of clusters •Key question is whether clusters are optimal and indicators are informative •Identifying optimal clusters for taxa or groups of taxa is possible using fidelity analysis, but may be tedious