Fishes of Texas Project Texas Natural History Collection · Fishes of Texas – past, present,...

Texas Natural History Collection Fishes of Texas Project

Not only data! A remarkably powerful

platform for the advancement of knowledge and research on the ecological health of streams and rivers

A model for open access, efficient delivery of data, and community driven data improvement

Texas Natural History Collection

Fishes of Texas – past, present, future

Fishes of Texas Project - overview

Advanced spatial ecology at landscape

scale invasive species models climate change impact conservation planning bioassessment

Museum specimen-based occurrence data

Such rigorous applications of data are not possible without major investment in compilation, processing and quality control

Texas Natural History Collection Fishes of Texas Project – data origins

42 museums represented

Our data come only from museum specimens (for now)


Track 1 Track 2

Dates of acquisition 11/1997 to 05/2006 04/2009 to 10/2010

Number of records 81,218 43,223 new (not including updates of

Track 1 records) Records georeferenced 67,190 ~40,000

Distinct georeferenced localities 5,729 + ~2,000 not in Track 1

Collecting events 10,954 ? Taxa synonymized with accepted names Yes mostly done

Geographic outliers flagged 3,789 (5%) none

Geographic outliers verified completed none

Collection dates edited 3,114 none

Collector names edited Yes none

fields available on website all verbatim data only, no GIS-based categorical fields

Date range 1851-2006 1853 - 2010

Number of contributors 34 9 new + 29 previous donor updates


Texas Natural History Collection Fishes of Texas Project

Georeferencing Methods (Manual vs. Automated) - Chose manual method (HerpNet Protocol)

• did not use lat/longs data obtained from data donors • no program could batch process and provide error radii • <2 years to finish - ~20,000 locations (synonymized to ~6,000)

- tested automated program (BioGeomancer)

• Only 27% of BG points fell within manually determined error radii • Median of 11km difference between BG points and manual

• (we wanted to work at 1 km for modeling, etc.) • In 73% of cases errors overlapped • In 52% of cases BG points were southeast of manual point

BioGeomancer Manual

able to georeference

60% of points that we manually

georeferenced 100% more than one possibility 14% 0%

median error radius 9 km 2 km


123,000 records 1852 – 2010 7,553 locations

Fishes of Texas

Project

Fishes of Texas Project – spatial coverage


Track 1 Track 2

Dates of acquisition 11/1997 to 05/2006 04/2009 to 10/2010

Number of records 81,218 43,223 new (not including updates of

Track 1 records) Records georeferenced 67,190 ~40,000

Distinct georeferenced localities 5,729 + ~2,000 not in Track 1

Collecting events 10,954 ? Taxa synonymized with accepted names Yes mostly done

Geographic outliers flagged 3,789 (5%) none

Geographic outliers verified completed none

Collection dates edited 3,114 none

Collector names edited Yes none

fields available on website all verbatim data only, no GIS-based categorical fields

Date range 1851-2006 1853 - 2010

Number of contributors 34 9 new + 29 previous donor updates


Texas Natural History Collection Fishes of Texas Project – discoveries

IDENTIFICATION ERROR RATES 1. Track 1 flagged records (geographic outliers) = 3789 (almost 5% of all records)

• 70% (2427) proved to be mis-identified (now corrected)

N lots examined

N lots with ID error percent error

Atractosteus spatula 2 2 100 % Dorosoma cepedianum 388 31 8 % Dorosoma petenense 165 24 15 % Erimyzon oblongus 32 0 0 % Erimyzon sucetta 42 2 5 % Fundulus jenkinsi 8 6 75 % Fundulus pulvereus 24 0 0 % Hybopsis amnis 21 5 24 % Ichthyomyzon castaneus 3 2 67 % Ichthyomyzon gagei 23 0 0 % Ictalurus furcatus 60 8 13 % Lepisosteus oculatus 35 8 23 % Lepisosteus osseus 46 14 30 % Lepisosteus platostomus 10 9 90 % Pomoxis annularis 49 2 4 % TOTAL 908 113 12 %

un-vouchered identifications should always be interpreted cautiously!!!!

2. How many that didn’t get flagged might be incorrectly identified? We can’t look at every jar of specimens, but we could relatively easily look at suspect “problem species pairs” in our own collection (TNHC):

Texas Natural History Collection Fishes of Texas Project – data cleaning


3. After ID verification

2. After georeferencing and taxonomic synonomization

1. Prior to data improvement

(coordinates provided by donors)

Hybognathus nuchalis

Creation of occurrence maps for 3 species

Micropterus punctulatus

Micropterus treculii Guadalupe Bass

Mississippi Silvery Minnow

Spotted Bass

Fishes of Texas Project – data cleaning


DISCOVERIES Ca

nadi

an

Red

Sulp

hur

Cypr

ess

Sabi

ne

Nec

hes

Trin

ity

San

Jaci

nto

Braz

os

Colo

rado

Colo

rado

-Lav

aca

Lava

ca-G

uada

lupe

Guad

alup

e

San

Anto

nio

Guad

alup

e-La

vaca

San

Anto

nio-

Nue

ces

Nue

ces

Rio

Gran

de

N NEW (33 total; 2.36/basin) 1 3 5 3 3 2 4 1 1 2 2 1 2 3Atractosteus spatula upstream reaches of Brazos and Trinity

Campostoma anomalumDorosoma petenense generally further west across state

Erimyzon sucetta X

Erimyzon oblongus X

Etheostoma radiosum X

Etheostoma proliare X

Etheostoma fusiforme X

Fundulus grandis X

Fundulus chrysotus X

Fundulus jenkinsi X

Heterandria formosa X

Hybognathus placitus X X X X

Ichthyomyzon gagei X X

Ictiobus cyprinellus X

Lepisosteus oculatus X

Lepisosteus platostomus X

Lepomis symmetricus X X

Lucania parva X

Lythrurus fumeus X

Moxostoma congestum X

Notropis amabilis X X

Notropis atherinoides X X X

Notropis chalybeaus X

Notropis straminius X

Phenacobius mirabilis X

Poecilia formosa X


Cana

dian

Red

Sulp

hur

Cypr

ess

Sabi

ne

Nec

hes

Trin

ity

San

Jaci

nto

Braz

os

Colo

rado

Colo

rado

-Lav

aca

Lava

ca-G

uada

lupe

Guad

alup

e

San

Anto

nio

Guad

alup

e-La

vaca

San

Anto

nio-

Nue

ces

Nue

ces

Rio

Gran

de

N NEW (33 total; 2.36/basin) 1 3 5 3 3 2 4 1 1 2 2 1 2 3Atractosteus spatula upstream reaches of Brazos and Trinity

Campostoma anomalumDorosoma petenense generally further west across state

Erimyzon sucetta X

Erimyzon oblongus X

Etheostoma radiosum X

Etheostoma proliare X

Etheostoma fusiforme X

Fundulus grandis X

Fundulus chrysotus X

Fundulus jenkinsi X

Heterandria formosa X

Hybognathus placitus X X X X

Ichthyomyzon gagei X X

Ictiobus cyprinellus X

Lepisosteus oculatus X

Lepisosteus platostomus X

Lepomis symmetricus X X

Lucania parva X

Lythrurus fumeus X

Moxostoma congestum X

Notropis amabilis X X

Notropis atherinoides X X X

Notropis chalybeaus X

Notropis straminius X

Phenacobius mirabilis X

Poecilia formosa X

New major river basin occurrence records

(Track 1 only)

Texas Natural History Collection Fishes of Texas Project - overview

• FISHES OF TEXAS DATABASE

Standardized

Verified

Fully Georeferenced

Quality Controlled Records with

metadata

Field notes

Images

Checklists

Keys & Species Accounts

Wealth of Data & Tools

Comment on anything

Submit photos

Submit field notes

Community Input

Texas Natural History Collection Fishes of Texas Project – now online

WEBSITE

http://www.fishesoftexas.org

Texas Natural History Collection Fishes of Texas Project – documentation

Texas Natural History Collection Fishes of Texas Project – website queries

Exte

nsiv

e co

ntex

tual

hel

p /

link

to m

etad

ata

Texas Natural History Collection Fishes of Texas Project – website mapping “s

uspe

ct sp

ecim

ens”

Texas Natural History Collection Fishes of Texas Project – taxonomic portal

Texas Natural History Collection Fishes of Texas Project – checklists

Texas Natural History Collection Fishes of Texas Project – species accounts

Texas Natural History Collection Fishes of Texas Project – website models

Texas Natural History Collection Fishes of Texas Project – keys

Texas Natural History Collection Fishes of Texas Project – image library In

situ

and

aqu

ariu

m

phot

os b

y Ga

rold

Sne

egas

273 illustrations by Joe Tomelleri

Texas Natural History Collection Fishes of Texas Project – digital library

Documents pertaining to individual specimens (or lots): • 1,384 jar contents photos • 1,081 specimen closeup photos • 229 x-rays



Images of species: • 220 illustrations (Tomelleri) (of 178 taxa) • 414 in situ photos (Sneegas) • 229 photos (freshly preserved with colors – Thomas & Bonner) • 12 3-D interactive animations (Labay)

Fishes of Texas Project – digital library



Images of species: • 220 illustrations (Tomelleri) (of 178 taxa) • 414 in situ photos (Sneegas) • 229 photos (freshly preserved with colors – Thomas & Bonner) • 12 3-D interactive animations (Labay)

1,199 Collectors’ field notes (pdfs)

Fishes of Texas Project – digital library


APPLICATIONS OF FOTX DATA & MODELS

http://www.fishesoftexas.org

Texas Natural History Collection Fishes of Texas Project – Species Distribution Models

Species Distribution Models provide a probability of occurrence of species for every square km statewide (Texas = 696,200 km2)


Junc

tion

Fred

ricks

burg

Fishes of Texas Project – bioassessment

Pragmatic testing of models – do they really work “on the ground”?


•Never before sampled (with voucher specimens) for fishes

•Hypothesized to be “one of the most pristine watersheds in Texas” (report to Environmental Defense)

•Part of South Llano Watershed Alliance (help with landowner contacts)

Junc

tion

Fred

ricks

burg



0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Modeled maximum probability of occurrence in the James River watershed for all modeled Colorado River basin species

Results of 2011 survey: collected Not (yet) collected


Colorado River basin fishes


0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1



Above this line species’ model predicts > 50% probability the species will be found in this basin < 50% probability below line



0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1




Normally uncommon /sporadic species



0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1




A priori suspected biotic interaction not accounted for in models - i.e. should expect one or other, but rarely both




0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1




A priori known issues with model quality A priori suspected biotic interaction not accounted for in models - i.e. should expect one or other, but rarely both




CONSERVATION PLANNING RESEARCH

Fishes of Texas Project – conservation planning


Input: 128 freshwater fish Species Distribution Models (SDMs) Objective: Identify a target representation of diversity using complementarity While simultaneously accounting for: 1. Connectivity (maximize) 2. # of cells (minimize) 3. # of clusters (minimize) 4. Shape (minimize edge)



Input: 128 freshwater fish SDMs (= 89,113,600 species probability values) Objective: 20% representation Base multi-criteria 1. Connectivity maximized 2. # of cells minimized 3. # of clusters minimized 4. Shape (edge minimized)


Texas Natural History Collection Fishes of Texas Project – bioassessment

BIOASSESSMENT RESEARCH


Labay B, Cohen AE, Sissel B, Hendrickson DA, Martin FD, et al. (2011) Assessing Historical Fish Community Composition Using Surveys, Historical Collection Data, and Species Distribution Models. PLoS ONE 6(9): e25145. doi:10.1371/journal.pone.0025145


Texas Natural History Collection Fishes of Texas Project - bioassessment

1. assessed historical condition

2. models provided a new benchmark (reference condition)

3. Compared to current conditions

= Bioassessment

Could this be done at a broader scale?

In Barton Creek we


Classical methods of bioassessment (e.g. IBI) have been severely criticized:

I. not transferable to large spatial scales,

II. not predictive assessments of deviation from natural states,

III. depend on reference sites, which have confounding

interactions with the first two goals above and are susceptible

to shifting baselines,

IV. not grounded in basic ecological understanding,

V. labor intensive and expensive to develop

Fishes of Texas Project - bioassessment

“Toward bioassessment without reference sites using species distribution modeling”

Texas Natural History Collection Fishes of Texas Project – bioassessment

Species A B C Site (km2) probability 0.87 0.19 0.09

Ecologically relevant Threshold ? ? ?

Prediction of Presence or

Absence P A A


Independent data from 269 sites sampled by 4 studies: • Studies differ in sampling

intensity

• compared observed communities (from surveys) to predicted (by models)



• 100 SDMs approximate habitat suitability INDEPENDENT of human impacts

•deviations from model predictions are a function of sampling quality, level of human impact, and model quality.


Independent data from 269 sites sampled by 4 studies: • Studies differ in sampling

intensity

• compared observed communities (from surveys) to predicted (by models)



• 100 SDMs approximate habitat suitability INDEPENDENT of human impacts

•deviations from model predictions are a function of sampling quality, level of human impact, and model quality.

~8 in/yr ~60 in/yr


y = 0.5342x + 0.9624 R² = 0.696

0

10

20

30

40

50

60

0 10 20 30 40 50 60

y = 0.2909x + 6.1245 R² = 0.3948

0

10

20

30

40

50

60

0 10 20 30 40 50 60

y = 0.5139x + 0.3078 R² = 0.4696

0

10

20

30

40

50

60

0 10 20 30 40 50 60

y = 0.2551x + 3.7392 R² = 0.3578

0

10

20

30

40

50

60

0 10 20 30 40 50 60

A

C

B

D

No.

obs

erve

d sp

ecie

s N

o. o

bser

ved

spec

ies

No. predicted species No. predicted species

TXST TCEQ

TPWD Tech

Relationship between the number of native fish species predicted and observed per site for A) the TXST dataset, B) the TCEQ dataset, C) the TPWD dataset, and D) the Tech dataset. The solid line represents equality of predicted and observed richness. Slopes are directly correlated with our assessment of the level of sampling intensity at each site for each survey (Tech highest, TXST, TCEQ, TPWD lowest).



25

30

35

40

45

50

55

60

0 0.2 0.4 0.6 0.8 1

IBI S

core

Observed/Predicted Ratio

Ecoregion 33 & 35 Ecoregion 32 Linear (Ecoregion 33 & 35) Linear (Ecoregion 32)

y = 19.84x + 37.531 R2 = 0.1636

20

25

30

35

40

45

50

55

60

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Aver

age

(200

6-20

08) I

BI sc

ore

Observed/Predicted Ratio

Ecoregion 29

Ecoregion 32

Ecoregion 33

y = 16.084x + 35.103 R² = 0.1685


The observed/predicted ratio for the TCEQ and TPWD datasets showed a significant though weak correlation with IBI scores. Not directly comparable though. Careful. Not to replace IBIs, but supplement their perspective.

TPWD

TCEQ


CONCLUSION

Model-based bioassessment appears promising:

• Addresses all criticisms of classical bioassessment

• EFFICIENT: create models once, assess deviations statewide with any survey dataset.

• GREAT IMPROVEMENT POTENTIAL: models can be improved by incorporating more validated occurrences and other relevant variables (hydrology!)



Research Overview: • Range Shifts & new discoveries • Conservation Planning • Bioassessment • Climate Change impact predictions

and much more is possible

All of this depends on: broadscale biodiversity specimen collections meticulous maintenance and improvement of the data

based on the specimens serving it, together with other relevant data, to

scientists, managers and the public and getting their input.

Fishes of Texas Project – the big picture


Dean Hendrickson

Ben Labay Doug Martin

Adam Cohen

Blak

e Si

ssel

Fishes of Texas Project – staff, collaborators, colleagues, volunteers


Val Bugh

Melissa Casarez Jeremy Harrison

Jessica R. Rains

Sahotra Sarkar

Bob Edwards

Gary Garrett (& wife Linda) Gina Higby

Mackenzie Anderson

Katharine Criswell (no photo yet)

Fishes of Texas Project – staff, collaborators, colleagues, volunteers


Allison Anderson (TNHC) - Database Jonathan Armbruster (AUM) - data donation Bob Ayers (SR) - Land access Henry Bart (TU) - data donation Mark Brinkman (SIB) - Collector Henry Brooks (None) - Land access Bill Bunch (SOS) - Advisor Sara Cartwright (OMNH) - data donation Melissa Casarez (TNHC) – cataloging, georeferenciing Ethan Cohen (TNHC) - Database advise Brandon Crawford (TNC) - Land access Drew Davis (TNHC) - General assistance Laura Dugan (TNHC) - GIS Anthony Echelle (OSU) - data donation Robert Edwards (UTPANAM) - Key - various critical tasks Hersh Eric (CRWR) - GIS data donation Gena Esposito (TNHC) - General assistance Lloyd Findley (CIAD) - Common names Margaret Fischer (TNHC) - Administrative Montemayor Gaby (CIAD) - Common names John Gallner (UTDIIA) - Database advise Gary Garrett (TPWD) – Funding - Key - various critical tasks Wendy Gordon (TCEQ) - Institutional liaison Keene Haywood (UTDIIA) - Website advise

Clark Hubbs (UT-IB) – started it all Ben Labay (TNHC) – Georeferencing, GIS Manuel Lemus (None) - Photographer Joann Lovelace (TNHC) - Georeferencing

John Lundberg (ANSP) - data donation John Maisano (TNHC) - Museum exhibits Edie Marsh-Matthews (OMNH) - data donation Floyd (Doug) Martin (TNHC) – Collector, determiner, etc. William Matthews (OMNH) - data donation Robby Maxwell (TNHC) - Collector Cydney Meyer (TNHC) - Collector George Murphy (TNHC) - Georeferencing Claire Patenia (TNHC) - Collector Susannah Reilly (TNHC) - Collector Nelson Rios (TU) - data donation

Jessica R. Rains (TNHC) - Cataloging- various critical tasks Stephen Ross (GCRL) - data donation Beck Runte (TNHC) - General assistance Sahotra Sarkar (UT – IB) – species occurrence modeliing Terri Siegenthaler (SR) - Land access Blake Sissel (UT – IB)– modeling Garold Sneegas (None) - Photographer Wayne Starnes (NCSM) - data donation Ann Syptak (TNHC) - General assistance Edward Theriot (TNHC) - Administrative Joe Tomelleri (None) - Illustrator Dora Wakou (TNHC) - Collector Jameson Wall (TNHC) - Collector Mark Westneat (FMNH) - data donation

Apologies to many (dozens of) volunteers and about 20 data providers not yet listed, and profound thanks to the 3,045 collectors who, over the past 160 years, deposited vouchers in Natural History Collections, thus assuring permanency and verifiability of these invaluable data. And, thanks to the institutions who assured the long-term maintenance of those collections.

Fishes of Texas Project – Acknowledgements

TEXAS ADVANCED COMPUTING CENTER

Texas Natural History Collection Fishes of Texas Project – THE END

www.fishesoftexas.org

Texas Natural History Collection Fishes of Texas Project – The End

Texas Natural History Collection Fishes of Texas Project – climate change

CLIMATE CHANGE RESEARCH


Notropis oxyrynchus –potential habitat may increase with climate change

PROJECTED CLIMATE-BASED HABITAT SUITABILITY SHIFTS Perspective of the direction and magnitude of climatic pressure

Fishes of Texas Project – climate change


~ 50,000 cataloged jars of preserved fishes

~ 47% (23,500 jars) from Texas freshwaters

~ 12% (6,000 jars) collected by TPWD

~ 17% (8,500 jars) collected by government agencies in general

All records verifiable via inspection of specimens.

TNHC Fish Collection – summary

TNHC Fish Collection

Texas Natural History Collection Fishes of Texas Project – what we do

Fishes of Texas Project Texas Natural History Collection · Fishes of Texas – past, present,...

Documents

Transcript of Fishes of Texas Project Texas Natural History Collection · Fishes of Texas – past, present,...