Justification


       The latest edition of Merrit et al. (2019) characterizes recent macroinvertebrate taxonomic revisions, many of which could have implications for multi-metric indices used for assessment purposes by the Pennsylvania Department of Environmental Protection (PADEP). Through collaboration with central office and regional staff, multiple taxa revisions have been identified that may affect index scores if revisions were to be adopted. An analysis is proposed to:


1. quantify at a general level how many monitoring sites/samples contain taxa that have been revised;
2. quantify the changes to macroinvertebrate multi-metric index scores at a sample-level due to taxonomic revisions; and
3. characterize the impact to method precision estimates.


       These objectives will be addressed throughout three Tiers of analysis.

Taxa Affected


       Below are tables of taxa that are affected by revisions, by revision type. Each table contains the following information:

  1. Group column indicates grouping of each taxonomic revision. For example if 4 genera of Ceratopogonidae are to be lumped, this represents one group.
  2. Importance column indicates the potential for changes to metric/index scores based on the abundance of the taxa and significance of the revision. 1 has the highest potential to changes scores, 3 the lowest.
  3. Old and new family/genus columns describe the taxonomic specifics of the revision.
  4. In SLIMS (old and new) columns indicate whether or not old/new taxa are currently in the ‘MI_BUG_CODES_LUT’ in SLIMS.
  5. Description and Implications columns give more detail regarding the change.
  6. Actions and BCG PTV analysis describe actions necessary to amend SLIMS tables to accomodate revision, and status of tolerance values for each taxa.
  7. Remaining PTV/BCG columns compare values for taxa prior to and after revisions, when available.


Taxonomic Revision Types:


1. New taxa:
  • Newly created genera/families.
  • There are 4 newly created groups of taxa proposed, including new genera in the Caenidae, Leptohyphidae, Pediciidae, and Baetidae families


Table 1. New Taxa: Newly created genera/families.
Group Type I Importance Order Old Family New Family Old Genus In SLIMS (old) New Genus In SLIMS (new) Type II Short Description Implications Actions BCG PTV analysis old PTV old BCG WARM old BCG COLD new PTV new BCG WARM new BCG COLD diff PTV diff BCG WARM diff BCG COLD
1 New 2 Ephemeroptera Caenidae Caenidae Sparbarus N New New genus in the Caenidae family Uncertain how common Add Sparbarus to SLIMS, all information needed Use Caenidae Family level PTV (7) 7 -7 0 0
1 New 2 Ephemeroptera Caenidae Caenidae Susperatus N New New genus in the Caenidae family Uncertain how common Add Susperatus to SLIMS, all information needed Use Caenidae Family level PTV (7) 7 -7 0 0
2 New 2 Ephemeroptera Leptohyphidae Leptohyphidae Ableptemetes N New Ableptemetes - new genus of Leptohyphidae MC: extreme southern US NA No family PTV in SLIMS 0 0 0
2 New 2 Ephemeroptera Leptohyphidae Leptohyphidae Allenhyphes N New Allenhyphes - new genus of Leptohyphidae MC: extreme southern US NA No family PTV in SLIMS 0 0 0
2 New 2 Ephemeroptera Leptohyphidae Leptohyphidae Homoleptohyphes N New Homoleptohyphes - new genus of Leptohyphidae Uncertain how common Add Homoleptohyphes to SLIMS, all information needed No family PTV in SLIMS 0 0 0
2 New 2 Ephemeroptera Leptohyphidae Leptohyphidae Leptohyphes Y New Leptohyphes - new genus of Leptohyphidae Uncertain how common NA Adopt PTV from MC & SLIMS for Leptohyphes (4) 4 4 0 0 0
2 New 2 Ephemeroptera Leptohyphidae Leptohyphidae Tricoryhyphes N New Tricoryhyphes - new genus of Leptohyphidae Uncertain how common Add Tricoryhyphes to SLIMS, all information needed No family PTV in SLIMS 0 0 0
2 New 2 Ephemeroptera Leptohyphidae Leptohyphidae Vacupernius N New Vacupernius - new genus of Leptohyphidae MC: extreme southern US NA No family PTV in SLIMS 0 0 0
3 New 2 Diptera Tipulidae Pediciidae Tricyphona N New Subgenus Tricyphona elevated to generic status in family Pediciidae (new) Tipulidae genera common, but uncertain how common these new genera will be Add Tricyphona to SLIMS, all information needed No family PTV 0 0 0
3 New 2 Diptera Tipulidae Pediciidae N New New family (Pediciidae) Add Pediciidae to SLIMS, all information needed No family PTV 0 0 0
4 New 3 Ephemeroptera Baetidae Baetidae Waynokiops N New Waynokiops is new genus in Baetidae (found mostly in lakes) Not here, not expected NA Use PTV from family Baetidae (6) 6 -6 0 0



2. Changes:
  • Taxa that have been assigned to new families/genera. The former family/genus no longer exists.
  • There are 4 taxonomic changes including two taxa placed into different genera, and sub-genera and sub-family taxa elevated to generic and family status, respectively.


Table 2. Changes: Taxa that have been assigned to new families/genera. The former family/genus no longer exists.
Group Type I Importance Order Old Family New Family Old Genus In SLIMS (old) New Genus In SLIMS (new) Type II Short Description Implications Actions BCG PTV analysis old PTV old BCG WARM old BCG COLD new PTV new BCG WARM new BCG COLD diff PTV diff BCG WARM diff BCG COLD
5 Change 2 Diptera Tipulidae Tipulidae Dolichopeza N New Subgenus Dolichopeza elevated to generic status in family Tipuilidae Tipulidae genera common, but uncertain how common these new genera will be Add Dolichopeza to SLIMS, all information needed Adopt family (Tipulidae) PTV (4) 4 -4 0 0
5 Change 2 Diptera Tipulidae Tipulidae Prinocera N New Subgenus Prinocera elevated to generic status in family Tipuilidae Tipulidae genera common, but uncertain how common these new genera will be Add Prinocera to SLIMS, all information needed Adopt family (Tipulidae) & MC PTV (4) 4 -4 0 0
6 Change 2 Crustacea Cambaridae Cambaridae Orconectes Y Faxonius N Change Orconectes is now Faxonius in family Cambaridae Orconectes found in ~2.5% of samples. No changes expected if TVs are consistent Add Faxonius to SLIMS, all information needed Adopt Orconectes PTV & BCG 6 4 4 6 4 4 0 0 0
6 Change 2 Crustacea Cambaridae Cambaridae Orconectes Y End Date Orconectes is now Faxonius in family Cambaridae Apply end date to Orconectes in SLIMS 0 0 0
7 Change 3 Trichoptera Glossomatidae Glossomatidae Matrioptila N Padunia N NA Matrioptila jeanae is now Padunia jeanae (Family Glossomatidae) Not here, not expected, Southeast US NA NA 0 0 0
8 Change 3 Trichoptera Uenoidae Thremmatidae Neophylax Y Neophylax Y Change Subfamily Thremmatinae is now a distinct family (Thremmatidae) and with its included genera Neophylax and Oligophlebodes removed from Uenoidae Neophylax common (~33% of samples), but assuming TVs are retained should not result in any changes NA Retain PTV & BCG 3 3 3 3 3 3 0 0 0
8 Change 3 Trichoptera Uenoidae Thremmatidae Oligophlebodes N Oligophlebodes N Change Subfamily Thremmatinae is now a distinct family (Thremmatidae) and with its included genera Neophylax and Oligophlebodes removed from Uenoidae Oligophlebodes not present in dataset Add Oligophlebodes to SLIMS, all information needed Need PTV & BCG. No family level in SLIMS
8 Change 3 Trichoptera Uenoidae Thremmatidae New Subfamily Thremmatinae in family Uenoidae is now a distinct family (Thremmatidae) Add Thremmatidae to SLIMS, all information needed Need PTV & BCG. No family level in SLIMS 0 0 0



3. Splits:
  • One to many – Taxa that were formerly part of a family/genus and are now assigned to multiple new/different family/genera. The former family/genus may or may not retain valid taxonomic status.
  • There are 9 taxa splits that have occurred. The most notable taxa splits (Importance = 1 or 2) include Polycentropus, Centroptilum, Paraleptophlebia, Cylindrotomidae, Gomphus, Sialis, and Cordulegaster.


Table 3. Splits: Taxa that were formerly part of a family/genus and are now assigned to multiple new/different family/genera.
Group Type I Importance Order Old Family New Family Old Genus In SLIMS (old) New Genus In SLIMS (new) Type II Short Description Implications Actions BCG PTV analysis old PTV old BCG WARM old BCG COLD new PTV new BCG WARM new BCG COLD diff PTV diff BCG WARM diff BCG COLD
9 Split 1 Trichoptera Polycentropodidae Polycentropodidae Polycentropus Y Holocentropus N Split Polycentropus is now divided into Holocentropus, Plectrocnemia, and Polycentropus Polycentropus is widespread (979/4202 samples; 23%) Add Holocentropus to SLIMS, all information needed Adopt family PTV in SLIMS (6) 6 4 4 6 0 4 4
9 Split 1 Trichoptera Polycentropodidae Polycentropodidae Polycentropus Y Plectrocnemia N Split Polycentropus is now divided into Holocentropus, Plectrocnemia, and Polycentropus n = 876 Small Free, 85 Large Free, 6 LS, 9 SWMMI, 3 MH Add Plectrocnemia to SLIMS, all information needed Adopt family PTV in SLIMS (6) 6 4 4 6 0 4 4
9 Split 1 Trichoptera Polycentropodidae Polycentropodidae Polycentropus Y Polycentropus Y Keep Polycentropus is now divided into Holocentropus, Plectrocnemia, and Polycentropus Potential for changes in diversity and sensitivity metrics depending on TVs NA Retain PTV from Polycentropus (genus) or Polycentropodidae (both 6) 6 4 4 6 0 4 4
10 Split 2 Ephemeroptera Baetidae Baetidae Centroptilum Y Anafroptilum N Split Baetidae Centroptilum ends. Now Anafroptilum and Neocloeon Centroptilum uncommon. Found in < 0.5% of samples Add Anafroptilum to SLIMS, all information needed Use PTV from MC (2) 2 3 3 2 0 3 3
10 Split 2 Ephemeroptera Baetidae Baetidae Centroptilum Y Neocloeon N Split Baetidae Centroptilum ends. Now Anafroptilum and Neocloeon Changes in MMI scores as a result expected to be minimal Add Neocloeon to SLIMS, all information needed Use PTV from Centroptilum (2) or Baetidae (6)? 2 3 3 2 3 3
10 Split 2 Ephemeroptera Baetidae Baetidae Centroptilum Y End Date Baetidae Centroptilum ends. Now Anafroptilum and Neocloeon End date for Centroptilum in SLIMS Use PTV from Centroptilum (2) or Baetidae (6)? 2 3 3 2 3 3 0 0 0
11 Split 2 Ephemeroptera Leptophlebiidae Leptophlebiidae Paraleptophlebia Y Neoleptophlebia N Split Paraleptophlebia split into Paraleptophlebia and Neoleptophlebia Paraleptophlebia found in ~43% of samples. Uncertain how common Neoleptophlebia will be Add Neoleptophlebia to SLIMS, all information needed Adopt PTV from MC (1) 1 2 2 1 0 2 2
11 Split 2 Ephemeroptera Leptophlebiidae Leptophlebiidae Paraleptophlebia Y Paraleptophlebia Y Keep Paraleptophlebia split into Paraleptophlebia and Neoleptophlebia Paraleptophlebia found in ~43% of samples. Uncertain how common Neoleptophlebia will be NA Adopt PTV from MC & SLIMS (1) 1 2 2 1 2 2 0 0 0
12 Split 2 Diptera Tipulidae Cylindrotomidae Phalacrocera Y Split Phalacrocera genus in Cylindrotomidae family (new) Uncertain how common NA Adopt PTV from SLIMS for Phalacrocera (4) 4 -4 0 0
12 Split 2 Diptera Tipulidae Cylindrotomidae Triogma Y Split Triogma genus in Cylindrotomidae family (new) Uncertain how common NA Adopt PTV from SLIMS for Triogma (4) 4 -4 0 0
12 Split 2 Diptera Tipulidae Cylindrotomidae N New Cylindrotomidae family (new) Uncertain how common Add Cylindrotomidae (family) to SLIMS, all information needed No family PTV 0 0 0
13 Split 2 Odonata Gomphidae Gomphidae Gomphus Y Phanogomphus N New Phanogomphus new genus in Gomphidae, Gomphus no longer exists Gomphus present in < 0.5% of samples. Uncertain how common new genera will be. Add Phanogomphus to SLIMS, all information needed Adopt family PTV (4) 5 4 4 4 3 3 1 1 1
13 Split 2 Odonata Gomphidae Gomphidae Gomphus Y Hylogomphus N New Hylogomphus new genus in Gomphidae, Gomphus no longer exists Gomphus present in < 0.5% of samples. Uncertain how common new genera will be. Add Hylogomphus to SLIMS, all information needed Adopt MC PTV (5) 5 4 4 5 0 4 4
13 Split 2 Odonata Gomphidae Gomphidae Gomphus Y Gomphurus N New Gomphurus new genus in Gomphidae, Gomphus no longer exists Gomphus present in < 0.5% of samples. Uncertain how common new genera will be. Add Gomphurus to SLIMS, all information needed Adopt MC PTV (5) 5 4 4 5 0 4 4
13 Split 2 Odonata Gomphidae Gomphidae Gomphus Y Stenogomphurus N New Stenogomphurus new genus in Gomphidae, Gomphus no longer exists Gomphus present in < 0.5% of samples. Uncertain how common new genera will be. Add Stenogomphurus to SLIMS, all information needed Adopt family PTV (4) 5 4 4 4 3 3 1 1 1
13 Split 2 Odonata Gomphidae Gomphidae Gomphus Y End Date Gomphus is no longer a valid genus in family Gomphidae Gomphus present in < 0.5% of samples. Uncertain how common new genera will be. End date for Gomphus in SLIMS 0 0 0
14 Split 2 Megaloptera Sialidae Sialidae Sialis Y Protosialis N Split Protosialis split from Sialis in family Sialidae Sialis found in ~6.6% of samples. Uncertain how common Protosialis will be Add Protosialis to SLIMS, all information needed Adopt MC PTV (4) 6 5 5 4 2 5 5
14 Split 2 Megaloptera Sialidae Sialidae Sialis Y Sialis Y Keep Protosialis split from Sialis (Sialis retained) in family Sialidae Sialis found in ~6.6% of samples. Uncertain how common Protosialis will be NA Retain PTV & BCG 6 5 5 6 5 5 0 0 0
15 Split 2 Odonata Cordulegastridae Cordulegastridae Cordulegaster Y Zorena N Split Zorena split from Cordulegaster in family Cordulegastridae Cordulegaster found in ~1.6% of samples. Uncertain how common Zorena will be Add Zorena to SLIMS, all information needed Adopt family PTV (3) 3 3 3 3 0 3 3
15 Split 2 Odonata Cordulegastridae Cordulegastridae Cordulegaster Y Cordulegaster Y Keep Zorena split from Cordulegaster (Cordulegaster retained) in family Cordulegastridae Cordulegaster found in ~1.6% of samples. Uncertain how common Zorena will be NA Retain PTV & BCG 3 3 3 3 3 3 0 0 0
16 Split 3 Ephemeroptera Polymitarcyidae Polymitarcyidae Tortopus N Tortopsis N Split New genus Tortopsis split from Tortopus Not here, not expected, Southern US NA Adopt PTV from family Polymitarcyidae (2) 2 2 0 0 0
16 Split 3 Ephemeroptera Polymitarcyidae Polymitarcyidae Tortopus N Tortopus N Keep Tortopus retained Not here, not expected, Southern US NA Adopt PTV from family Polymitarcyidae (2) 2 2 0 0 0
17 Split 3 Odonata Libellulidae Libellulidae Ladona Y Libellula Y Split Ladona and Libellula are now separate genera in family Libellulidae Niether present in dataset NA Adopt MC/retain SLIMS PTV (8) 6 8 -2 0 0
17 Split 3 Odonata Libellulidae Libellulidae Ladona Y Ladona Y Split Ladona and Libellula are now separate genera in family Libellulidae Niether present in dataset NA Retain SLIMS PTV (6) 6 6 0 0 0



4. Lumps:
  • Many to one – Taxa that were formerly part of a multiple families/genera, that are now considered a singular taxonomic group.
  • There were 8 groups of taxa lumped together. The most notable lumps inlcude Ceratopsyche - Hydropsyche, Maccaffertium - Stenonema, Promoresia- Optioservus, and Ceratopogonidae taxa.


Table 4. Lumps: Taxa that were formerly part of a multiple families/genera, that are now considered a singular taxonomic group.
Group Type I Importance Order Old Family New Family Old Genus In SLIMS (old) New Genus In SLIMS (new) Type II Short Description Implications Actions BCG PTV analysis old PTV old BCG WARM old BCG COLD new PTV new BCG WARM new BCG COLD diff PTV diff BCG WARM diff BCG COLD
18 Lump 1 Trichoptera Hydropsychidae Hydropsychidae Ceratopsyche Y Hydropsyche Y Lump Ceratopsyche is now a synonym of Hydropsyche (Family Hydropsychidae). This lump was examined by D Shull previously Already evaluated, but both taxa remain in SLIMS. Found together in 3.5% of samples NA Retain PTV from Ceratopsyche, same as Hydropsyche 5 4 4 5 5 5 0 -1 -1
18 Lump 1 Trichoptera Hydropsychidae Hydropsychidae Ceratopsyche Y Hydropsyche Y Lump Ceratopsyche is now a synonym of Hydropsyche (Family Hydropsychidae). This lump was examined by D Shull previously End date for Ceratopsyche in SLIMS OR change STR_ID_LEVEL to Hydropsyche. 5 4 4 5 5 5 0 -1 -1
19 Lump 1 Ephemeroptera Heptageniidae Heptageniidae Maccaffertium Y Stenonema Y Lump Maccaffertium collapsed into Stenonema again Large implications. Maccaffertium present in ~63% of samples. M. Cole believes this may not stick. NA Adopt PTV from MC and SLIMS for Stenonema (4) 3 3 3 4 4 4 -1 -1 -1
19 Lump 1 Ephemeroptera Heptageniidae Heptageniidae Maccaffertium Y Stenonema Y End Date Maccaffertium collapsed into Stenonema again IF ADOPTED: Maccaffertium needs end-date in SLIMS 0 0 0
20 Lump 1 Coleoptera Elmidae Elmidae Promoresia Y Optioservus Y Lump Promoresia lumped into Optioservus Promoresia found in ~17% of samples. NA Adopt PTV from MC & SLIMS for Optioservus (4) 2 2 3 4 4 4 -2 -2 -1
20 Lump 1 Coleoptera Elmidae Elmidae Promoresia Y Optioservus Y End Date Promoresia lumped into Optioservus Add end date to Promoresia in SLIMS Adopt PTV from MC & SLIMS for Optioservus (4) 0 0 0
21 Lump 2 Diptera Ceratopogonidae Ceratopogonidae Ceratopogon Y Subfamily - Ceratopogoninae N Lump Ceratopogonidae family genera are now instead lumped into the Ceratopogoninae subfamily >1 Ceratopogoninae taxa found in ~1.2% of samples. Uncertain how common new genera will be Change ‘STR_ID_LEVEL’ to Ceratopogoninae for Ceratopogon; also need to address Ceratopogonidae All have same PTV and BCG values 6 4 4 6 0 4 4
21 Lump 2 Diptera Ceratopogonidae Ceratopogonidae Probezzia Y Subfamily - Ceratopogoninae N Lump Ceratopogonidae family genera are now instead lumped into the Ceratopogoninae subfamily >1 Ceratopogoninae taxa found in ~1.2% of samples. Uncertain how common new genera will be Change ‘STR_ID_LEVEL’ to Ceratopogoninae for Probezzia; also need to address Ceratopogonidae All have same PTV and BCG values 6 4 4 6 0 4 4
21 Lump 2 Diptera Ceratopogonidae Ceratopogonidae Bezzia Y Subfamily - Ceratopogoninae N Lump Ceratopogonidae family genera are now instead lumped into the Ceratopogoninae subfamily >1 Ceratopogoninae taxa found in ~1.2% of samples. Uncertain how common new genera will be Change ‘STR_ID_LEVEL’ to Ceratopogoninae for Bezzia; also need to address Ceratopogonidae All have same PTV and BCG values 6 4 4 6 0 4 4
21 Lump 2 Diptera Ceratopogonidae Ceratopogonidae Stilobezzia Y Subfamily - Ceratopogoninae N Lump Ceratopogonidae family genera are now instead lumped into the Ceratopogoninae subfamily >1 Ceratopogoninae taxa found in ~1.2% of samples. Uncertain how common new genera will be Change ‘STR_ID_LEVEL’ to Ceratopogoninae for Stilobezzia; also need to address Ceratopogonidae All have same PTV and BCG values 6 4 4 6 0 4 4
22 Lump 3 Trichoptera Leptoceridae Leptoceridae Ylodes N Triaenodes Y Lump Ylodes is now a synonym of Triaenodes (Family Leptoceridae) Ylodes not present, Triaenodes rare (~0.5% of samples). NA Retain PTV & BCG from Triaenodes in SLIMS 6 3 3 -6 -3 -3
23 Lump 3 Ephemeroptera Baetidae Baetidae Pseudocloeon N Labiobaetis Y Lump Pseudocloeon lumped into Labiobaetis (family Baetidae) Labiobaetis already changed in SLIMS. Uncommon (< 0.3% of samples) NA Retain PTV & BCG from Labiobaetis 4 3 3 -4 -3 -3
24 Lump 3 Ephemeroptera Ephemerellidae Ephemerellidae Dentatella N Eurylophella Y Lump Ephemerellidae Dentatella lumped into Eurylophella Dentatella not present. Eurylophella relatively common (~20% of samples). NA Retain PTV & BCG from Eurylophella 4 2 3 -4 -2 -3
25 Lump 3 Ephemeroptera Heptageniidae Heptageniidae Nixe Y Afghanurus N Lump Nixe lumped into Afghanurus Nixe uncommon (< 0.7% of samples). Afghanurus not present. No changes assuming TVs consistent. Add Afghanurus to SLIMS, all information needed Adopt PTV from MC (2) 2 1 1 2 0 1 1
25 Lump 3 Ephemeroptera Heptageniidae Heptageniidae Nixe Y Afghanurus N End Date Nixe lumped into Afghanurus End Date for Nixe in SLIMS 0 0 0


Click here to download a spreadsheet containing the tables shown above



Analytical Methods



      Macroinvertebrate samples were pulled from SLIMS and imported into an Rstudio (3.6.3) project. Data were characterized by method, and samples collected using freestone (6D 200), limestone (2D 300), or multihabitat (10D 200) protocols were retained. Data were then joined to the NHDplus HiRes NHD Flowline by the COMID field to get drainage area for each segment associated with the data collection. This allowed separation of 6D 200 into small (<50 mi2) or large (>50 <1000 mi2) freestone, or SWMMI (>1000 mi2) methods. Samples that did not have latitude or longitude were retained for analysis but were not included in the maps generated below, or in precision estimate analysis (see Tier III) due to the uncertainty of location.


Tier I


      Tier I analysis was intended to give a general, high-level of idea of how prevalent changes due to taxonomic revisions are across PA. First, the percentage of samples containing one or more taxa that are subject to revision was calculated. Additional analysis was directed at determining how often taxa that are proposed to be lumped have been identified in the same sample in the past. Specifically, I determined how often Maccaffertium and Stenonema were present in the same samples. The same thing was done for taxa that are proposed to be lumped into the Ceratopogoninae subfamily (Ceratopogon, Probezzia, Bezzia, Stilobezzia), for Coleopterans Promoresia and Optioservus, and for Trichopterans Ceratopsyche and Hydropsyche.


Tier II


       After the sample dataset was categorized by method, metrics and overall scores were generated using the taxa collected in each sample. A list of taxa that are subject to revisions was generated along with associated Hilsenhoff tolerance values. Then, the samples were rescored to quantify the effects of taxa updates on individual metrics and overall multimetric scores.

       The relationship between IBI score before revisions and the difference in IBI scores that resulted from revisions was also examined. This was intended to give a general sense of whether sites scoring higher or lower were more or less affected.


Tier III


       The effect of taxonomic updates on precision estimates (PEs) of various methods was of interest. First, sample clusters were identified that were within 100 m and located on the same COMID. Samples collected through time in each cluster were used to calculate temporal PEs of each method with an adequate sample size. Temporal PEs include four sources of variability: natural intrasite, methodological, natural temporal, and variability due to changes in condition.

       In addition, samples within clusters that were collected on the same day were used to calculate intrasite PEs of each method with an adequate sample size. Intrasite PEs include are comprised of intrasite and methodological variability.



Results

Tier I


       The resulting dataset contained 93,905 rows of taxa from 4,202 samples collected between 1999 and 2019. The greatest number of samples were collected using the small and large freestone methods (Figure 1). Of the 4,202 samples, 777 did not have latitude or longitude. These samples were retained for analysis but were not included in the maps.


Figure 1. Total number of samples collected, by method, that were included in analysis.


Figure 2. Interactive map of 3,425 macroinvertebrate samples with included in the analysis with coordinate information. Points are symbolized by method and selectable by layer. Click each point for a pop-up box containing additional information.
* Basemap layers (topographic and aerial) can be changed interactively. Select the layer containing the method of interest, then click points to view additional information including stream name, drainage size, and COMID.


       A total of 3,841 out of 4,202 (91.4%) samples contained ≥1 taxa that are subject to revision. The taxa that are subject to revision most commonly contained in samples were Maccaffertium, Paraleptophlebia, and Polycentropus (Figure 3).


Figure 3. Percent occurence of taxa subject to revision.


       Since Maccaffertium and Stenonema are proposed to be lumped, the the number of samples that contained both taxa was quantified. These taxa were found together in 91 samples (2.2%), most often in small and large freestone samples. Maccaffertium was most often present in higher abundances within each sample, but the proportions of each taxa varied in each sample (Figure 4).


Figure 4. Number of Maccaffertium and Stenonema individuals in samples where both taxa were present, by method.


       Additionally, 4 taxa that are currently classified in the Ceratopogonidae family (Ceratopogon, Probezzia, Bezzia, Stilobezzia) are now instead lumped into the Ceratopogoninae subfamily. Out of the 4,202 sites in the dataset, ≥ 1 of these taxa were found together in 52 samples (1.2%). The small freestone method most often had these taxa congeners present together (Figure 5).


Figure 5. Number of Ceratopogon, Probezzia, Bezzia, Stilobezzia individuals in samples where ≥ 1 taxa were present, by method.


       Another taxa lump that was examined were two Coleopterans in the Elmidae family (Promoresia and Optioservus) that are now lumped into Optioservus. Out of the 4,202 sites in the dataset, these taxa were found together in 483 samples (11.5%). The small freestone method most often had these taxa congeners present together. Median abundance of Optioservus was higher for samples collected using freestone and limestone methods, while median abundance of Promoresia was higher for SWMMI samples (Figure 6). The multihabitat method only had 1 sample (out of 32 total) with both taxa present.


Figure 6. Number of Promoresia and Optioservus individuals in samples where both taxa were present, by method. Number of samples (n) is shown by method.


       The final, major taxa lump that was examined were two Trichopterans in the Hydropsychidae family (Ceratopsyche and Hydropsyche) that are now lumped into Hydropsyche. Out of the 4,202 sites in this dataset, these taxa were found together in 147 samples (3.5%). The small freestone method most often had these taxa congeners present together. Samples collected using the limestone and multihabitat methods did not have any samples with both taxa present. Median abundance of Ceratopsyche was higher for samples collected using freestone methods, while median abundance of Hydropsyche was higher for SWMMI samples (Figure 7).


Figure 7. Number of Ceratopsyche and Hydropsyche individuals in samples where both taxa were present, by method. Number of samples (n) is shown by method.

Click here to download a spreadsheet of taxonomic data from samples where the following taxa proposed to be lumped were present together:
  • Maccaffertium and Stenonema
  • Ceratopogon, Probezzia, Bezzia, Stilobezzia (≥2 taxa found together)
  • Promoresia and Optioservus
  • Ceratopsyche and Hydropsyche


Tier I Implications


       The vast majority (91.4%) of samples included in this analysis contained ≥ 1 taxa that are subject to revision. The implications for changes to IBI score vary depending on the nature of the revision, specifically if it is a new taxa, change, lump, or a split.


1. New taxa:
  • The impact of new taxa cannot be quanitatively examined. The occurrence and abundance of these taxa are not known in existing datasets using different taxonomic status.
    • Only Leptohyphes is currently present in SLIMS, and there were no occurences in the dataset examined (n = 4202 samples).


2. Changes
  • The effects of taxonomic changes can often be examined by a simple comparison of PTV’s before and after the revision. Some changes are more complex (e.g. subgenus to genus revision) that cannot be examined.
    • 1465 / 4202 samples (34.9%) had taxa subject to a change. These samples were mostly comprised of Neophylax (33%); many fewer Orconectes (2.5%)
    • If Neophylax and Orconectes maintain PTV and BCG values, no changes will result in MMI scores.


3. Splits
  • Overall, 2348 / 4202 samples (56%) of samples had >= 1 taxa subject to a split. Most were collected via the small freestone method (2102), followed by the large freesone method (184) (Table 5).
    • Taxa subject to splits that were most commonly included in samples: Paraleptophlebia (43%), Polycentropus (23%), and Sialis (7%)
  • The impact of taxa splits could only be examined with a very limited dataset. Archived samples containing taxa subject to splits (n = 17) were reexamined using updated Merritt and Cummins taxonomic keys.
    • Split taxa included in the limited dataset included:
      • Centroptilum
      • Cordulegaster
      • Paraleptophlebia
      • Polycentropus
      • Sialis


Table 5. Number (n) of samples that had >= 1 taxa subject to a split, by method. Percent (%) column represents percent of total dataset (n=4202).
Method n %
Small Freestone 2102 50.0%
Large Freestone 184 4.3%
SWMMI - Summer 23 0.5%
Multihabitat 15 0.4%
SWMMI - Fall 14 0.3%
Limestone 10 0.2%


4. Lumps
  • Lumps are most easily examined. The many to one nature will decrease the diversity and richness metrics. Additionally, if Hilsenhoff tolerance values are not equal for the taxa that are to be lumped, tolerance metrics may change.
    • 757 / 4202 samples (18%) had >= 1 taxa subject to a lump. Most were collected via the small freestone method (548), followed by the large freesone method (126) (Table 6).
      • The percentage of samples containing both taxa proposed to be lumped were highest for
        • Promoresia and Optioservus (11.5%)
        • Ceratopsyche and Hydropsyche (3.6%)
        • Maccaffertium and Stenonema (2.2%)
        • Ceratopogon, Probezzia, Bezzia, Stilobezzia (at least 2 in 1.2%)
    • One effect could be decreased diversity, another could be a change in PTV or BCG values if the taxa proposed to be lumped are not equal


Table 6. Number (n) of samples that had >= 1 taxa subject to a lump, by method. Percent (%) column represents percent of total dataset (n=4202).
Method n %
Small Freestone 548 13.0%
Large Freestone 126 3.0%
SWMMI - Summer 36 0.9%
Multihabitat 22 0.5%
SWMMI - Fall 21 0.5%
Limestone 4 0.1%


       There were 4, 4, 9, and 8 groups of new taxa, changes, splits, and lumps, respectively. Tier II analysis investigates the effect of taxa changes/lumps on individual metrics and overall multimetric index scores, which reveals additional insight into the effect size of taxonomic revisions. Quantitative analysis in Tiers II and III provides greater clarity.



Tier II

       After all taxonomic revisions were applied, IBI scores were re-calculated and compared to IBI scores prior to revision. Negative differences in scores reflect lower IBI scores after revisions were applied.

       The summer and fall SWMMI methods were most affected by taxonomic revisions, with median score differences of -5.6 and -5.0, respectively. In extreme cases, IBI scores could vary by 20 points or more. Median differences were comparatively minor for the remaining methods, but near double-digit differences were observed in a few samples (Figure 8).


Figure 8. Histograms of differences in IBI scores, by method. Negative values indicate IBI scores were higher before taxonomic revisions were applied.


Table 7. Median and range of differences in IBI scores before and after taxonomic revisions were applied, by method. Rows are in descending order of median difference.
Method n median range
SWMMI - Summer 222 -5.6 -25 - 0
SWMMI - Fall 143 -5.0 -23.3 - 0
Large Freestone 510 -1.5 -13.4 - 0
Limestone 64 -1.3 -3.6 - 0
Small Freestone 2892 -0.3 -9.7 - 11.2
Multihabitat 21 0.0 -2.4 - 0


       A total of 66 samples showed a >= 10 point difference in IBI scores as a result of taxonomic revisions. These samples were collected using the summer SWMMI (n = 45), fall SWMMI (n=19), and large freestone (n=2) methods. The taxa present in these samples that are subject to revision were Ceratopsyche, Optioservus, and Maccaffertium (Figure 9). The latter was most prevalent and responsible for the majority of scoring differences.


Figure 9. Bar plot of abundance of revised taxa in samples with a >= 10 point difference in IBI score, by method. Bars are filled with colors that represent the abundance of revised taxa in each sample. The black point on each bar indicates the absolute value of the change in IBI score as a result of taxonomic revisions.


       Taxa splits were difficult to evaluate because of their one to many nature. Further, most of the splits were new and previous identifications did not provide information on splits. We re-identified archived samples (n = 17) where Centroptilum, Cordulegaster, Paraleptophlebia, Polycentropus, and/or Sialis individuals were present. Re-identified samples may have changed an identification, but there were no samples where multiple taxa subject to a split were present.

       After re-identification, samples were rescored and compared to IBI scores that included lumps. There were no differences in IBI scores that resulted from splits.


       An interactive map was created to allow users to examine differences in IBI scores at the sample level before and after taxonomic revisions, for each method. Samples that included coordinates (n = 3,132) are symbolized by the magnitude of their differences in IBI scores, where darker sites have larger differences.


Figure 10. Interactive map of 3,132 macroinvertebrate samples included in the analysis with coordinate information. Points are symbolized by the magnitude of difference in IBI score before and after taxonomic revisions. Each method is a selectable layer that can be turned on and off. Click each point for a pop-up box containing additional information, including a taxa list before and after revisions were applied.
* Basemap layers (topographic and aerial) can be changed interactively. Select the layer containing the method of interest, then click points to view additional information including IBI scores before and after (and the difference), location information, and taxa lists before and after revisions were made.


       Next, the relationship between IBI score before revisions and the difference in IBI score due to taxonomic revisions was examined. This could inform whether sites scoring poorly or well are most often affected in a consistent direction. The relationship was examined visually and no statistical analysis was performed. Results show that there is a general negative relationship in SWMMI samples, where sites scoring better most often experience decreases in IBI score as a result of revisions. There is a good amount of variability, however. This pattern is observed to a lesser extent in the small and large freestone methods. Sites scoring poorly were rarely affected by taxonomic revisions, while sites scoring >= 50 were more often decreased after revisions were applied. There was no relationship between IBI score before and differences after in the limestone and multihabitat methods, although the sample size was much smaller (Figure 11).


Figure 11. Scatter plot of IBI score before revisions (x axis) vs. the difference in IBI scores resulting from revisions (y axis), by method.



Tier II Implications

  • Taxonomic revisions resulted in decreases in IBI scores across the board for all methods.

  • The taxonomic revision responsible for the greatest change in IBI scores was the Maccaffertium lump into Stenonema. The difference in tolerance values between these two taxa (Maccaffertium = 3 vs. Stenonema = 4), coupled with the ubiquitous nature of the taxa (Maccaffertium present in ~63% of samples in the dataset) caused tolerance and diversity based metrics to decrease after the revision.

    • Maccaffertium individuals are widespread, but their presence in SWMMI samples had a greater effect (decrease) on IBI scores compared to other methods.
    • When a taxa has different tolerance values than the taxa it is lumped into (e.g. Maccaffertium), both taxa do not have to be present in a sample to change IBI scores.
    • This lump is viewed with skepticism by Mike Cole and due to it’s outsized effect on IBI scores, is worth careful consideration.

  • Additional lumps, including Promoresia and Optioservus, Ceratopsyche and Hydropsyche, and Ceratopogon, Probezzia, Bezzia, Stilobezzia also cause decreases in IBI scores, but not to the same extent as Maccaffertium.

    • A 2013 analysis by D. Shull using a separate dataset (available below) found that in general there was a 1-2 point decrease in IBI scores due to lumping Ceratopsyche and Hydropsyche.
Click here to download Dustin Shull's presentation of his Ceratopsyche analysis


  • Splits had no effect on IBI scores, based on the limited analysis of re-identified samples herein.
    • The 17 samples examined for splits did not include multiple taxa. In the instances where taxonomy changed as a result of the split, tolerance values were carried forward from the parent taxa. Therefore, there was no impact to tolerance or diversity based metrics.

  • In general, IBI scores were decreased at SWMMI sites with better initial IBI scores.
    • The same was true for large and small freestone sites; however the magnitude of decreases was much smaller.


  • The effect of new taxa and changes cannot be examined until future taxonomic data become available. In the interim, taxa that are likely present in the Commonwealth that are not already present in SLIMS need to have records created, complete with tolerance and BCG values, trophic status, and bug codes.



Tier III

       After grouping samples spatially and temporally, intrasite and temporal PE’s were calculated for samples before and after taxonomic revisions were applied. Summer and fall SWMMI samples were excluded from PE calculations since data collection preceded development and final implementation of the method.

       Enough data were available to calculate intrasite PE’s for the small freestone method, and temporal PE’s for the small and large freestone and limestone methods. The influence of taxonomic revisions on PE’s was small and comparable to those included in assessment method documentation:

  • Freestone Intrasite: 5.4
  • Freestone Temporal: 10
  • Limestone Intrasite: 7.3
  • Limestone Temporal: 11.4


Table 8. Intrasite and temporal precision estimates (PE’s) before and after taxonomic revisions were applied, by method. Sample size of distinct spatial and/or temporal clusters (n_cluster) and number of samples within each cluster (n_samples) is included.
Method n_cluster n_samples PE_type PE_before PE_after
Small Freestone 29 60 intrasite 4.8 5.0
Large Freestone 3 6 intrasite
Small Freestone 141 405 temporal 9.0 9.0
Large Freestone 94 237 temporal 11.5 11.0
Limestone 18 36 temporal 12.5 12.2


Tier III Implications

  • Taxonomic revisions had little effect on PE’s for all methods with adequate sample sizes to perform calculations.

  • All PE’s calculated from this dataset were comparable to those published within each assessment method.

    • The effect on PE’s is a relatively minor concern compared to the differences in IBI scores that result from taxonomic revisions.




Acknowledgments

K. Bardell, M. Brickner, M. Lookenbill, and D. Shull provided input to guide the analysis. A. Blascovich, M. Brickner, T.Daley, and D. Rebuck re-identified samples to inform the analysis of taxonomic splits.