Appendices to the Bioindicators for Assessing Ecological Integrity of Prairie Wetlands
Bioindicators for Assessing Ecological Integrity of Prairie Wetlands
Report # EPA/ 600/ R-96/ 082
September 1995
Table Of Contents:
| Appendix A Appendix B Appendix C Appendix D Appendix E |
Appendix F Appendix G Appendix H Appendix I Appendix J |
Appendix K Appendix L Appendix M Appendix N Appendix O |
Appendix A.
Taxonomic Index: Plant Species Tolerances and Responses to Water Regime, Drainage, Land Use, Salinity, and Turbidity; Food Value to Waterfowl; Toxicity Data Availability
This data base primarily includes vascular plant species that the literature reports to be dominant or abundant in at least one prairie wetland; it does not include all prairie wetland plant species. Three aquatic mosses/liverworts and one alga are also included because of their large size. Blanks in this data base indicate a lack of information for the species. Codes in the data base mean
Form (from Reed 1988):
A = annual,
B = biennial,
P = perennial
N = native,
I = introduced
E = emergent
F = forb
G = grass
GL = grasslike
H = partly woody
H2 = horsetail
J = algae
M = aquatic moss
P3 = pepperwort
S = shrub
Z = submersed
$ = succulent
/ = floating
Dependence (from Reed 1988, Hubbard et al. 1988); Reed (1988) notes that the dependence category should not be equated strictly to degrees of wetness, because many obligate species occur in temporary or seasonal wetlands (although most occur in permanent or semipermanent wetlands).
OBL = Obligate wetland species that under natural conditions occur almost always (> 99% probability) in wetlands.
FACW = Facultative wetland species that usually occur in wetlands (67-99% probability) but occasionally are found in nonwetland upland (terrestrial) habitats.
FAC = Facultative species equally likely to occur in wetlands or nonwetland upland habitats.
FACU = Facultative upland species that usually occur in nonwetlands (67-99% probability) but occasionally are found in wetlands.
+, - = Slightly wetter (+) or drier (-) than indicated by one of the four acronyms above.
WaterType (from Kantrud et al. 1989):
SS = seasonally flooded wetlands
SP = semipermanently flooded wetlands
P = permanently flooded wetlands
T = temporarily flooded wetlands
CAPs = dominant in this habitat
Lower case = less dominant but occurs frequently
Drainage
i = Seeds are relatively intolerant of sustained drainage; these are the species that Galatowitsch (1993) found in natural wetlands but not in restored (reflooded) wetlands that had been drained for long periods.
t = Seeds are relatively tolerant of sustained drainage; these are the species whose seeds Wienhold and van der Valk (1989) found to be viable even after 30 years of drainage.
Land Use (from Kantrud et al. 1989):
g = Species typically invades grazed wetlands, or it disproportionately unaffected by grazing.
h = Species typically invades hayed wetlands, or it is disproportionately unaffected by haying activities.
p = Species typically invades plowed wetlands, or it is disproportionately unaffected by soil tillage.
r = Species typically occurs mainly in relatively undisturbed (reference) wetlands
Salinity (mainly from Kantrud et al. 1989):
f = fresh (< 800 uS conductivity, or < 3 g/L salt)
o = oligohaline (800-8,000 uS conductivity, or 4-20 g/L salt)
m = mesohaline (8,000-30,000 uS conductivity)
p = polyhaline (30,000-45,000 uS conductivity)
e = euhaline or hyperhaline (> 45,000 uS conductivity, or > 70 g/L salt)
* = documented specifically by literature (other entries were based on judgement of Kantrud et al. 1989)
CAPS = salinities where it is dominant (not merely occurring) according to Kantrud et al. 1989.
Turbidity (mainly from Kadlec and Wentz 1974, Davis and Brinson 1980, Nichols 1984a, Chambers and Kalff 1985):
t = Species typically invades wetlands with highly turbid water (minimal light penetration) or is disproportionately unaffected by turbidity increases.
x = Species typically occurs in wetlands with the least turbid water (great light penetration) or is highly sensitive to turbidity increases.
DuckFood? (mainly from Kadlec and Wentz 1974):
s = Seeds are frequently consumed by waterfowl.
f = Foliage and/or tubers are frequently consumed by waterfowl.
* = Highly preferred by some species of waterfowl.
RecsPhyto
The number of records for this species in USEPA's PHYTOTOX data base as of March 1994. Each record represents the response of the species to one substance during one investigator's experiment. G= number of records for the genus (not necessarily this species). The data base is presently in the process of being joined with USEPA's AQUIRE data base.
RecsAquire
The number of records for this species in USEPA's AQUIRE data base as of October 1993. Each record represents the response of the species to one substance during one investigator's experiment. The data base can be publicly accessed, and it provides a quantitative report of each experiment and the citation to the source literature.
Appendix B.
Taxonomic Index: Invertebrate Tolerances and Responses to Water Regime, Oxygen, Salinity, and Sediment; Food Value to Waterfowl
For each of > 160 invertebrate taxa, this data base describes what is known about preference or tolerance with regard to water regime, oxygen, salinity, and turbidity. The next-to-last column indicates taxa known to be consumed frequently by waterfowl. This data base does not include all invertebrate taxa found in prairie wetlands. Taxa are listed in phylogenetic order (as indicated by the Sequence field), and were included in this data base if (a) by numbers or weight they constituted a large portion of samples collected for a published study or for an unpublished data base that was made available, and/or (b) literature indicated a strong association with the particular environmental variable. As is evidenced by the many blanks in the data base, much information is unavailable for many of the included taxa; however, certain abbreviations and sources of information were used:
Reproduction (Repro) (mainly from Wiggins et al. 1980):
1 = Group 1--Overwintering Residents. Capable of passive dispersal only. Aestivate and overwinter in the dry basin either as drought-resistant cysts/eggs or as juveniles and adults.
2 = Group 2--Overwintering Spring Recruits. Reproduce in springtime surface water before it disappears because egg-laying depends on water. Aestivate and overwinter in the dry basin mainly as eggs or larvae (or for a few beetles, as adults).
3 = Group 3--Overwintering Summer Recruits. Can reproduce even when basin is dry because egg-laying does not require presence of surface water. Overwinter as eggs or larvae within the egg matrix.
4 = Group 4--Nonwintering Spring Migrants. Reproduce in springtime surface water before it disappears because egg-laying depends on water. Adults of the subsequent generation(s) leave the wetland as it dries and overwinter in permanent wetlands.
Water Regime (mainly from Swanson et al. 1974, LaBaugh and Swanson 1988, Driver 1977, Neckles et al. 1990, Bataille and Baldassarre 1993):
t = Occurs regularly in temporary and seasonal wetlands, as well as (usually) in semipermanent and permanent wetlands.
sp = Occurs predominantly in semipermanent and permanent wetlands, but may also be present in temporary and seasonal wetlands.
Oxygen (mainly from Beck 1977, Hilsenhoff 1982, Rosenberg and Resh 1993):
1 = Most tolerant of oxygen deficits, as commonly occurs with severe eutrophication.
4 = Least tolerant of oxygen deficits.
a = Numeric rating to which this code is appended refers to a broader or narrower taxonomic level to which this taxon belongs, and its applicability to this particular taxon is unknown.
b = Numeric rating to which this code is appended refers to taxa that appear to be closely related, and its applicability to this particular taxon is unknown.
Salinity (mainly from Timms et al. 1986, Timms and Hammer 1986, Lancaster and Scudder 1987):
f = fresh (< 800 µS/cm conductivity, or < 3 g/L salt)
o = oligohaline (800-8,000 µS/cm conductivity, or 4-20 g/L salt)
m = mesohaline (8,000-30,000 µS/cm conductivity)
p = polyhaline (30,000-45,000 µS/cm conductivity)
e = euhaline or hyperhaline (> 45,000 µS/cm conductivity, or > 70 g/L salt)
Sediment
t = Species typically invades wetlands with highly turbid water (minimal light penetration) or is disproportionately unaffected by turbidity increases.
x = Species typically occurs in wetlands with the least turbid water (great light penetration) or is highly sensitive to turbidity increases.
Duck Food?
Literature from prairie wetlands, cited in Sheehan et al. 1987, indicates major use by the following waterfowl species (H = hens, Y = young):
American Wigeon (AmWi), Blue-winged Teal (BwTe), Canvasback (Canv), Gadwall (Gadw), Lesser Scaup (LeSc), Mallard (Mall), Northern Pintail (Pint), Northern Shoveler (Shov), Redhead (Redh), Ruddy Duck (RuDu)
Sequence
This code is used to place the taxa in their approximate phylogenetic sequence. The first digit to the left of the decimal distinguishes among taxa in different phyla or orders, whereas the first digit to right distinguishes among orders or families, the second one to the right among families or genera, and so on.
Appendix C.
Taxonomic Index: Bird Wetland Type Associations, Relative Abundance, and Trends
This data base includes bird species that occur regularly--generally annually--at some season in multiple prairie wetlands and seem to use wetlands to a greater degree than nonwetland upland or deepwater habitats. This judgement was the author's, and it is based on reviewed literature (e.g., Duebbert 1981, Faanes 1982, Kantrud and Stewart 1984, Short 1989) and knowledge of the life history of species.(1)
The data base contains these columns:
AOU
A numeric code assigned by the American Ornithological Union, used to sort species into their approximate phylogenetic sequence.
Status (Migration, Breeding) (Status_mig, Status_br)
Information in this column on relative abundance is from Faanes and Stewart (1982) and applies to North Dakota; it is generally but not exactly applicable to other areas of the prairie region. Abundance of some species varies between spring vs. fall migration; in this data base, only the larger abundance term was used. Terms are as defined by Faanes and Stewart (1982):
abundant = very large numbers and easily observed
common = large numbers
fairly common = fair to moderate numbers
uncommon = low numbers
rare = very low numbers, but occurs somewhere at least annually
* = locally more numerous than indicated by the term
Type of Wetland (Wet_type)
The type of wetland (as defined by water regime) in which the species occurs most often. Inform ation in this data base was interpreted from Provost 1947, Weber et al. 1982, Faanes 1982, Kantrud and Stewart 1984, Hop et al. 1989, Colwell and Oring 1990, and Short 1989. Although nearly all of the listed species occur in nearly all wetland types, only the wetland types used most often are noted. Seasonal/interannual variation, geographic variation, differences between foraging and nesting preferences, and needs for multiple types are not accounted for. All species listed as occurring in temporary wetlands are also found along the margins of seasonal, semipermanent, and permanent wetlands. Where information allows, wetland types used to a greater degree are denoted by upper case.
a = alkali basins
p = permanently flooded basins
sp = semipermanently flooded basins
ss = seasonally flooded basins
t = temporarily flooded basins
Layers
The general layers of habitat in which the species generally is found. Information in this data base was interpreted from Provost (1947), Weber et al. (1982), Faanes (1982), Colwell and Oring (1990), and Short (1989). Although it is recognized that nearly all of the listed species can occur in nearly all layers, only the layers used most often are noted. Seasonal/interannual variation, geographic variation, and differences between foraging and nesting preferences, are not accounted for.
es = inhabits emergent vegetation growing out of hydric soil (surface water generally is absent)
ew = inhabits emergent vegetation whose stems are immersed in standing water
ow = inhabits patches of open water that are unvegetated except for presence of submersed aquatic plants
m = inhabits generally unvegetated mud, sand, bare soil, and gravel
t = inhabits trees or shrubs surrounding wetlands
Phenology
Actual peak breeding times vary somewhat depending on weather during a particular year. A numeric code indicates the usual peak breeding period for the species in North Dakota:
1 = 24 April to 7 June
2 = 14 May to 10 July
3 = 22 May to 19 July
Total Pairs (Pairs_tot)
The sum of breeding pairs found in 1993 in 416 North Dakota wetlands visited by Igl and Johnson (unpub. data, NPSC, Jamestown, ND). All wetlands included in this tally are in the Prairie Pothole Region.
Number of Wetlands (Num_Wets)
The number of wetlands where the species was present, based on visits by Igl and Johnson (unpub. data, NPSC, Jamestown, ND). Numbers are just for 1992-93 in North Dakota wetlands within the Prairie Pothole Region. Although data were collected primarily during the breeding season, these frequencies were calculated for migrant and breeding individuals combined.
Frequency in Wetlands (Freqwets)
The percentage of the wetlands where the species was present, based on visits by Igl and Johnson (unpub. data, NPSC, Jamestown, ND). Frequencies are just for 1992-93 in North Dakota wetlands within the Prairie Pothole Region. Although data were collected primarily during the breeding season, these frequencies were calculated for migrant and breeding individuals combined.
Maximum per Wetland (Max_per_w)
The largest number of breeding pairs counted in any single one of the 452 North Dakota prairie wetlands visited by Igl and Johnson (unpub. data, NPSC, Jamestown, ND).
Region of Widest Distribution (Reg_frqma)
The North Dakota subregion in whose wetlands in 1993 the species was most widely distributed, based on Igl and Johnson (unpub. data, NPSC, Jamestown, ND). Frequencies were calculated for migrant and breeding individuals combined. See Stewart and Kantrud (1972a) for a map of subregions.
Region of Greatest Mean Abundance (Reg_abumax)
The North Dakota subregion in whose wetlands the species had the greatest mean abundance (on a per wetland basis) during 1993, and based on Igl and Johnson (unpub. data, NPSC, Jamestown, ND). The mean abundances were calculated for migrant and breeding individuals combined.
Region (Bbs_region)
This is a marker for data in the columns that follow to the right.
E = Data for the eastern part of the prairie region, including the Red River Valley in eastern North Dakota and extending north slightly into Manitoba, all of western Minnesota and nearly all of southern Minnesota (including some nonprairie areas), and north-central Iowa (corresponds to Breeding Bird Survey [BBS] stratum #40).
C = Data for the central part of the prairie region, including much of eastern South Dakota, eastern and central North Dakota, a small part of southern Manitoba, and southcentral Saskatchewan and Alberta (corresponds to BBS stratum #37).
W = Data for the western part of the prairie region, including central South Dakota, central and northwestern North Dakota, northern Montana, and southern Saskatchewan and Alberta (corresponds to BBS Stratum #38).
# of Routes (Bbs_numrts)
The number of BBS routes in the region on which the species has ever been found. Approximately 135 routes were run in the region at least once during the period 1966-1991, which is the period from which the data in this data base came. Each route consists of 50, 3-minute stops along a 25-mile roadside route. The routes encompass all habitats, not just wetlands.
Pct of Routes (Bbs_rts)
The percentage of BBS routes on which the species has ever been found in the specified region.
# per Route (Bbs_avg_rt)
For the specified region, this is the number of individuals of the species found on the average along a BBS route.
Maximum Frequency (Bbs_maxrt)
For the specified region, this is the greatest frequency of occurrence (% of 50 stops) found along any BBS route.
Priority (Priortybb)
For the specified region, this is the trend in frequency of occurrence amongst BBS routes during the period 1966-1991. Because of uneven coverage among routes and years during the period, a bootstrapping technique (Sauer and Droege 1990) was used by the USFWS in calculating the trends. The author identified categories with these priorities:
1 = The frequency with which the species was encountered declined on more routes than it increased during the period, and the difference was statistically significant.
2 = The frequency with which the species was encountered declined on more routes than it increased during the period, but the difference was not statistically significant.
3 = The frequency with which the species was encountered increased on more routes than it decreased during the period, but the difference was not statistically significant.
4 = The frequency with which the species was encountered increased on more routes than it decreased during the period, and the difference was statistically significant
? = The species may breed in the subregion but has not been detected along BBS routes.
Appendix D.
Taxonomic Index: Published Field Studies of Plant-Salinity Relationships in Prairie Wetlands
This data base indexes studies that have reported on prairie wetland plant associations with salinity. It primarily includes species that the literature reports to be dominant or abundant in at least one prairie wetland; it does not include all prairie wetland plant species. A few of the included species are normally "upland" species (Reed 1988) even though they were found in a wetland.
Appendix E.
Taxonomic Index: Published Field Studies of Plant-Water Regime Relationships in Prairie Wetlands
This data base indexes studies that have reported on prairie wetland plant associations with various water regimes. It primarily includes species that the literature reports to be dominant or abundant in at least one prairie wetland; it does not include all prairie wetland plant species. A few of the included species are normally "upland" species (Reed 1988) even though they were found in a wetland.
Appendix F.
Taxonomic Index: Published Field Studies of Invertebrate--Vegetation Cover Relationships in Prairie Wetlands
This data base indexes studies that have reported on invertebrate associations with various vegetation types, densities, and patterns in prairie wetlands. It primarily includes invertebrates that the literature reports to be dominant or abundant in at least one prairie wetland; it does not include all prairie wetland invertebrate species.
Appendix G.
Taxonomic Index: Published Field Studies of Invertebrate--Water Regime Relationships in Prairie Wetlands
This data base indexes studies that have reported on invertebrate associations with various water regimes in prairie wetlands. It primarily includes invertebrates that the literature reports to be dominant or abundant in at least one prairie wetland; it does not include all prairie wetland invertebrate species.
Appendix H.
Dominant Algae in Prairie Wetlands
This data base includes algae (and a few Protista) that the literature reports to be dominant or abundant in at least one prairie wetland; it does not include all prairie wetland algal species.
Appendix I.
Rare Wetland Plants Reported From Prairie Counties of North Dakota
The list of rare plants with county distributions was provided by Douglas Eiken, North Dakota Natural Heritage Inventory, April 1991; it has been edited to include only species officially considered to be wetland dependent according to Reed (1988). These codes are assigned to the data:
Form (from Reed 1988):
A = annual, B = biennial, P = perennial
N = native, I = introduced
E = emergent
F = forb
G = grass
GL = grasslike
H2 = horsetail
Z = submersed
$ = succulent
/ = floating
Dependence (from Reed 1988):
Reed (1988) notes that the dependence category should not be equated strictly to degrees of wetness, because many obligate species occur in temporary or seasonal wetlands (although most occur in permanent or semipermanent wetlands). These abbreviations have been used:
OBL = Obligate wetland species that under natural conditions occur almost always (> 99% probability) in wetlands.
FACW = Facultative wetland species that usually occur in wetlands (67-99% probability) but occasionally are found in nonwetland upland (terrestrial) habitats.
FAC = Facultative species equally likely to occur in wetlands or nonwetland upland habitats.
FACU = Facultative upland species that usually occur in nonwetlands (67-99% probability) but occasionally are found in wetlands.
+, - = Slightly wetter (+) or drier (-) than indicated by the acronym.
Appendix J.
Catalog of Published Biological Studies of Prairie Wetlands: Locations, Sampling Regimes, Key Variables, and Related Descriptors
This data base lists virtually all published studies since about 1970 that reported collection of multispecies biological data from at least one prairie wetland. Abbreviations are interpreted as follows:
IndepVar
A list of the major independent variables that were measured by the study. Some variables (e.g., temperature, season) that are not primarily anthropogenic were not included in this table. The abbreviations are as follows:
h = hydrologic gradient or fragmentation (among sites or within a site)
H = hydrologic manipulation (depth, duration, etc.)
n = nutrient gradient (N, P)
N = nutrient manipulation/dosing
p = pesticide or heavy metal gradient
P = pesticide or heavy metal dosing
s = salinity gradient (or alkalinity, Ca, Mg, SAR)
S = salinity manipulated
t = turbidity, light penetration, or sedimentation--gradient
T = turbidity, light penetration, or sedimentation--manipulated
c = cover density or cover ratio--gradient
C = cover density or cover ratio--manipulated
#Yrs
The maximum number of years covered by the study; some treatments or collections that are a part of the study may have covered fewer years.
LocState1...3
This column indexes the state(s) or province(s) where data were collected.
LocCo1...10
This column indexes the counties where data were collected; a few studies covered more than 10 counties.
LocName
This column provides more specific information describing the location.
The succeeding columns first provide information on bird studies, then invertebrate studies, vascular plant studies, and amphibian studies, as follows:
BirdGrp
The group of birds that were surveyed:
B = all birds
f = waterfowl (ducks, geese, swans)
F = one species of waterfowl
G = one species of gamebird
s = shorebirds (sandpipers, plovers, etc.)
w = waterbirds (mostly nonpasserine aquatic species)
BirdVars
The measurements that were reported:
b = biomass, weight, or standing crop
f = frequency, numbers, or density (of individuals, by species)
j = duration of use
k = home range size
m = mortality
p = production or growth
r = richness
t = trace or heavy metal content
v = behavior
BirdWets
The maximum number of major sample units (e.g., wetlands, fields) that were surveyed.
BirdSeas
Season(s) covered by the survey:
p = spring
s = summer
f = fall
w = winter
BirdFreq
Average frequency of surveys:
d = daily
w = weekly (< w means more often than weekly but not daily)
bw = biweekly
t = triweekly
m = monthly
2 = (the number of visits per year)
InvVars
The measurements reported:
b = biomass or weight or standing crop (by taxon)
B = biomass or weight or standing crop (all taxa combined)
f = frequency or numbers or density (of individuals, by taxon)
m = mortality
n = nutrient or caloric content
r = richness
InvWets
The maximum number of major sample units (e.g., wetlands) that were surveyed.
InvReps/Wet
The maximum number of samples or transects per wetland.
InvSeas
Season(s) covered by the survey:
p = spring
s = summer
f = fall
w = winter
InvFreq
Maximum frequency of sample collections:
w = weekly (< w means more often than weekly but not daily)
bw = biweekly
m = monthly
2 = (the number of visits per year)
InvIDlevel
Most specific level to which most organisms were identified:
f = family (fc = only Coleoptera, fm = only midges)
o = order
s = species
VPlantVars
Same codes as for InvVars, plus:
a = alkaline (mono)phosphatase activity
b = biomass or weight or standing crop
c = percent cover by species (C= not broken out by species)
d = decomposition rate
e = electronic transfer system (ETS) flux
f = frequency, numbers, or stem density (of individuals, by taxon)
F = total number or density (not broken out by taxon)
g = germination rate by species (G= all species combined)
h = height or length
m = mortality
n = nutrient or calorific content
p = production or growth
r = richness
s = seed production
t = trace or heavy metal content
u = glucose mineralization or metabolism
x = oxygen consumption or respiration
VPlantWets
See InvWets above.
VPlantReps/Wet
See InvReps/Wet above.
VPlantSeas
See InvSeas above.
Algae
Similar to VPlant columns.
Amph
Similar to VPlant columns.
Appendix K.
Catalog of Ongoing Biological Studies of Prairie Wetlands: Locations, Sampling Regimes, Key Variables, and Related Descriptors
This data base lists many, but surely not all, studies that are currently underway in the prairie region and involve collection of multispecies biological data from at least one prairie wetland. Included are funded studies whose field work has not been completed, recent studies whose field work has been completed but whose data have not been analyzed completely for publication, and recent studies whose results have been reported so far only as abstracts at symposia. Information fields include the investigator's name, phone number, bibliographic reference if any, starting year (YRSTART), projected completion year (YREND), number of wetlands studied (WETS), major variables examined (see Appendix J abbreviations), and location.
Appendix L.
Descriptions of Data Sets Analyzed For This Report
Below are described the unpublished data sets that were analyzed and used to derive estimates of variance reported in Appendices M and N as well as the species accumulation reported in Appendix O.
I. INVERTEBRATES
A. Hanson Activity Traps
Records are indexed by year, period (month), wetland, sample, and vegetation. Each record represents a unique combination of year-period-wetland-sample-vegetation for certain metrics: number of species, number of individuals (all species combined), biomass (all species combined). Only the data from wetlands that lacked high densities of fathead minnows were included. Data are from activity traps and cover two years. In year one, there were 5 sample periods, 3 wetlands sampled per period, and 8 random samples per wetland. In year two, there were 4 sample periods (3 identical to year 1), the same 3 wetlands (plus one other) were sampled, and 10 random samples were collected per wetland. Vegetation condition (present/absent) of each sample was recorded incidentally and was not part of the sampling design. In both years, data were collected using the same equipment and mostly the same protocol. The total number of unique records is 320. The data are from wetlands in east Polk County in west-central Minnesota; they were provided by Dr. Mark Hanson, Minnesota Dept. of Natural Resources, Bemidji, Minnesota.
B. Euliss Sweep Nets
Records are indexed by year, wetland, transect, and sample. Each record represents a unique combination of year-wetland-transect-sample for certain metrics: number of species, number of individuals (all species combined). Samples were collected monthly with a sweep net, but data are not broken out by month; only season totals were available. Data are from two years. In year one, 16 wetlands were sampled; there were 2-8 transects per wetland with 1-2 samples per transect for a total of 118 unique records. In year two, the same 16 wetlands plus two new ones were sampled, using 5 transects per wetland with 2-3 samples per transect for a total of 265 unique records. The data were provided by Dr. Ned Euliss, NPSC, Jamestown, ND.
C. Euliss Sediment Traps
Records are indexed by region, plot(2) , wetland(3) , transect, wetland class, and health class. Data are from 36 wetlands, all sampled once in 1993. Each record represents a unique combination of region-plot-wetland-transect-wetlandclass-healthclass, for certain metrics: number of individuals (all species combined), biomass (all species combined). There are 2 regions, 35 wetlands, 3 wetland classes (semipermanent, seasonal, temporary), 10 plots, 5 transects, and 2 health classes. There are 18 wetlands in each of the two regions, 8-10 plots per wetland class, and 5 transects per plot. For each of the two health classes ("healthy" and "unhealthy"), there are approximately equal numbers of plots, wetlands, transects, wetland classes, and regions. Total number of unique records is 180. Data represent an entire growing season's collection of settled, decay-resistant remains of snails, cladocerans, ostracods, and clam shrimp, as measured using a standard protocol. The data were provided by Dr. Ned Euliss, NPSC, Jamestown, ND.
D. MERP Substrate Samplers
Records are indexed by year, period (month), zone (cover type), and depth class. Data are from 2-6 wetlands, but unfortunately the wetland identifier was lost and all the data from individual wetlands were combined. Each record represents a unique combination of year-period-zone-depth class, for certain metrics: number of species, number of individuals (all species combined), biomass (all species combined). Data cover five years. There are 21-24 periods per year, 6-7 zones per period, and 2 depth classes per zone. There are 7 zones per year and 18-25 periods per zone. Total unique records is 790. In all years, data were collected using the same equipment and a standardized protocol. The data are from experimentally manipulated wetland cells (mesocosms) of the Marsh Ecology Research Program (MERP) located in the Delta Marsh, Manitoba. The data were provided by Dr. Henry Murkin, Institute for Wetland and Waterfowl Research, Oak Hammock Marsh,
Manitoba.
E. MERP Activity Traps
Records are indexed by year, period (month), wetland ("cell"), treatment, and zone. Each record represents a unique combination of year-period-wetland-treatment-zone, for the following metrics: number of species, number of individuals (all species combined), biomass (all species combined). Data cover five years (except biomass data, which exclude 1988). There are 23-24 periods per year, 1-6 wetlands per period, 1 treatment per wetland, and 2-6 zones per treatment. There are 20-25 periods per zone and 17-23 perio ds per wetland. There are 2-6 wetlands per zone and the same number per treatment. There are 1-4 zones per wetland and 6-7 zones per treatment. Total unique records is 1200. In all years, data were collected using the same equipment and a standardized protocol. Data are highly variable because conditions among wetlands were intentionally and dramatically manipulated as part of experiments. The data are from experimentally manipulated wetland cells (mesocosms) of the Marsh Ecology Research Program (MERP) located in the Delta Marsh, Manitoba. The data were provided by Dr. Henry Murkin, Institute for Wetland and Waterfowl Research, Oak Hammock Marsh, Manitoba.
F. Duffy Data Set
Records are indexed by wetland and period (sampling date). Each record is the mean of four samples per wetland per sample period, and it represents a unique combination of wetland-period for certain metrics: number of species, number of individuals (all species combined), biomass (all species combined). Data are all from a single year, representing four wetlands: 2 wetlands were sampled for four periods, 1 wetland for six periods (including the same four periods), and one for nine periods. Total unique records is 23. In all periods and wetlands, data were collected using the same equipment (a 3.5-inch PVC corer) and a standardized protocol. The data are from Deuel County, South Dakota, and they were provided by Dr. Walter Duffy, South Dakota Cooperative Fish and Wildlife Research Unit, Brookings, South Dakota.
II. BREEDING BIRDS
A. FWS Breeding Bird Survey
Records are indexed by year, state, and route. Each record represents a unique combination of year-state-route, for these metrics: number of species, number of individuals (all species combined), number of sample points containing wetland species). Routes are surveyed repetitively (but often noncontinously) among years, so that a partly different set of routes constitutes each year's data. All data were collected using a standardized protocol. There are 21 years, 5 states (and provinces), and 51 routes. When routes are combined, there are 95 unique year-state combinations. Other characteristics are
years per route: 1-21
routes per year: 19-30
routes per state: 2-16
routes per state per year: 1-14
years per state: 14-21
states per year: 5
The total number of unique records is 543; the data were provided by Sam Droege, National Biological Service, Washington, D.C.
B. Igl & Johnson Data Set
Records are indexed by year, wetland, and visit. Each record represents one wetland, described in term of these metrics: number of breeding species, number of breeding + nonbreeding species, number of breeding pairs, number of breeding plus nonbreeding individuals. The maximum value from both visits and both years was used. Visits were made to 480 wetlands, 330 in 1992 and 416 in 1993. Each was visited 1-2 times per year, and visits were timed to cover both early and late breeders. The sampling design is as described by Stewart and Kantrud (1972). The study covered nearly all of North Dakota, but only the data from sites in the Prairie Pothole Region were used in this analysis (i.e., Missouri Slope and Little Missouri Slope subregion data were excluded).
Appendix M.
Results of Power of Detection Analyses of Existing Prairie Data Sets
This table describes the results of the components of variance approach that was applied to estimate sample power of detection. Methods of calculation are described on page 15. Results are reported under these column headings:
Taxon
The major group for which variability was determined.
Metric
A measured characteristic of the biological community that was reported by the original investigator.
Units
The units of measurement (e.g., grams) in which the metric is expressed.
Equipment
Sampling equipment used by the collector of the original data.
RandomVar
The variable for which the number of requisite samples and the associated level of precision were calculated.
Data set
The source of the data; the experimental design associated with the analyzed data is detailed in Appendix L.
Pr_forS10
The specified metric's precision that would be obtainable if 10 samples of the type described were collected, e.g., a sample size of 10 would allow the user to distinguish a difference between two means of [specified] grams, given the assumptions of the analysis. This figure is based on the "optimistic" equation, but in most instances the value from the "conservative" equation differed only slightly (see p. 15 for discussion of the equations). "Ten samples" was an arbitrary number selected so that relative levels of resultant precision could be roughly compared among taxa, sampling methods, and sampling designs.
BreakPt
The number of samples beyond which increasing the number of samples results in relatively little increase in the level of precision. This number was estimated visually and subjectively from plotted curves, so the true value could be plus or minus about two samples.
DetectDif1
An example of the detectable difference in two means. This example is for the smallest difference that can be predicted from the data we used (i.e., the "optimistic" estimate, as described on p. ?).
SSizeMin1
The smallest number of samples that would be able to detect the specified difference, based on the "optimistic" calculation method defined on page 15.
SSizeMax1
The smallest number of samples that would be able to detect the specified difference, based on the "conservative" calculation method defined on page 15.
DetectDif2
A second example of the detectable difference in two means. This example is for the largest difference that can be predicted from the data we used (i.e., the "conservative" estimated as described on p. ?).
SSizeMin2, Max2
As above, but for the second data set.
PooledVars
Variables that were subsumed in defining the samples (records) that were used for the analysis.
Appendix N.
Library of Coefficients of Variation from Prairie Wetlands
This data base quantifies the biological variability of prairie wetlands at multiple spatial and temporal scales using a standard statistic--the coefficient of variation--which is defined as:
CV = (100 * SD) / ,
where SD is the standard deviation of the data, is the mean, and CV is the coefficient of variation.
The 412 coefficients were calculated from 15 prairie wetland data sets, encompassing both published literature and raw data provided to the author by investigators in the region. The data's quality and representativeness of prairie wetlands in general are unknown. Data were obtained on a purely opportunistic basis, and it is not known how much of the variability described by any CV can be attributed to ecological variation as opposed to human-related effects.
This data base was developed to illustrate the relative degree of spatial and temporal variability that can be expected in a given situation, recognizing that each situation is unique; therefore, the CVs must be interpreted with caution. Comparisons among CVs are confounded by the fact that the sampling designs that resulted in each CV were not necessarily well-balanced (e.g., for a given study, not all habitats were sampled equally, at the same times of year, etc.). More significantly, and especially when values were extracted from the literature, it was seldom clear what data might have been previously combined (e.g., replicates composited) in arriving at a particular mean, standard deviation, or CV. When it was apparent to what a particular mean was referring (i.e., which variables had been combined and/or averaged), data then were referenced accordingly in the data base. CVs were always reported at the finest level of detail (least amount of pooling) possible as well as for pooled samples. The least-pooled sample CVs are indicated by presence of an "x" in the AllSamp column, coinciding with absence of an "x" or "a" from all remaining columns.
These abbreviations are used in the data base:
Group
The broad taxonomic group that was the primary focus of the study, namely,
b = birds
i = invertebrates
p = plants
SampType
The type of equipment or protocol used:
emtrap = emergence trap
actrap = activity trap
quad = quadrat
depth 0-5 cm = seeds collected from sediment depths of 0-5 cm
Metric
biomass = collective weight of all individuals
numindiv = the number of individuals
numtaxa = the number of taxa (richness)
seed den = density of seedlings
shoots/m2 = density of plant shoots/m2
Regions
Samples from different subregions of the prairie region:
x = The CV was calculated after pooling data from multiple regions.
a, A = The CV represents the among-region variability (the number following this letter indicates the number of pooled samples).
37 = The CV is only for the central part of the prairie region.
38 =The CV is only for the western part of the prairie region
40 = The CV is only for the eastern part of the prairie region
HealthCl
Samples from the two "wetland health" classes defined by the cited study:
x = The CV was calculated after pooling data from both health classes.
a, A = The CV represents the variability between the two health classes.
States
Prairie states or provinces covered by the BBS data
x = The CV was calculated after pooling data from multiple states.
a, A = The CV represents the among-states variability after all routes and years within a state were pooled.
Routes
The BBS route data (each route is 25 mi long and contains 50 stops where data are collected):
x = The CV was calculated after pooling data from multiple routes.
a, A = The CV represents the among-route variability after each route's years were pooled.
WetTypes
Data by wetland types:
x = The CV was calculated after pooling data from multiple wetland types.
a, A = The CV represents the among-type variability.
When the CV is just for the specified wetland type:
p = permanently flooded
ss = seasonal
sp = semipermanent
t = temporary
Wets
Data by discrete wetland:
x = The CV was calculated after pooling data from multiple wetlands.
a, A = The CV represents the among-wetland variability.
When the CV represents variability just within a single wetland, a number has been arbitrarily assigned a reference number (1,2, etc. are arbitrarily assigned reference numbers).
Treat
Data by wetland treatment regime:
x = The CV was calculated after pooling data from multiple wetlands, each having been hydrologically manipulated in a different manner.
a, A = The CV represents the among-treatment variability.
When the CV represents variability just within a single treatment regime:
d = Wetlands were subjected to drawdown.
f = Wetlands were subjected to flooding.
h = wetlands in which emergent plants were harvested.
t1, t2 = Two different treatment regimes.
Trans
Data by transects within wetlands:
x = The CV was calculated after pooling data from multiple transects within the wetlands.
a, A = The CV represents the among-transect variability.
Zones
Data by vegetation zones within wetlands:
x = The CV was calculated after pooling data from multiple zones within a wetland.
a, A = The CV represents the among-zone variability.
When the CV represents variability just within the specified zone, a zonal term has been included: wet meadow, shallow, etc. (zonal terms are mostly those of the originator of the data)
Depths
Data by depth class within wetlands:
x = The CV was calculated after pooling data from multiple depth zones within a wetland.
a, A = The CV represents the among-depth zone variability; the two zones were < 30 cm and > 30 cm.
Veg
Data by vegetated condition within wetlands:
x = The CV was calculated after pooling data from both vegetated and unvegetated (open water) areas of the wetland.
a, A = The CV represents the between-area variability.
AllSamps
x = The CV was calculated after pooling data from multiple wetlands, transects, zones, months, years, etc., unless other columns are blank (see explanation in second example query below).
Yrs
Data grouped by year:
x = The CV was calculated after pooling data from multiple years.
a, A = The CV represents the among-year variability.
When the CV represents variability just within a single yea, a number has been to distinguish discrete years (1, 2, etc. are arbitrarily assigned numbers).
Months
Data grouped by month or biweekly sampling period:
x = The CV was calculated after pooling data from multiple sampling periods.
a, A = The CV represents the among-sampling period variability,
When the CV represents variability just within a single sampling period, numbers or a season abbreviation were assigned arbitrarily to distinguish discrete periods (1,2, etc.):
es = early summer
ls = late summer
p = spring
CV
The coefficient of variation calculated for samples grouped in the manner defined by the preceding columns
CVmin
The smallest of several coefficients of variation calculated for samples grouped in the manner defined by the preceding columns.
CVmax
The largest of several coefficients of variation calculated for samples grouped in the manner defined by the preceding columns.
The organization and structure of the data base are best illustrated by two example queries, as follows:
Query #1
How much does species richness (number of taxa) vary among wetlands?
Approach:
Find (sort(4)) all records that have an entry "numtaxa" (number of taxa) in the column "Metric" AND have an entry preceded by a lower-case "a" in the column "Wets" (among wetlands). In this example, there are 12 records meeting these criteria. Inter-wetland variability is quantified by the corresponding values in the column "CV." Different CV values for different records reflect other influencing conditions. In this example, one CV was calculated for each of six sampling periods indexed in the Months column. Four wetlands (indicated by "a4" in the Wets column) were sampled on May 9 and 31, June 14, and June 29 (indicated in the Months column) and the among-wetland CVs calculated for each date. An "x" in the AllSamps column indicates that individual samples from within each wetland had been pooled prior to calculating the CV. In addition, another "a4" record lacks any date in the Months column, but rather has an "x." This indicates a situation where, for each wetland in the data set, the data from all the months were pooled before calculating the CV. The CVmin column then reports the smallest among-month CV from the four wetlands and the CVmax column reports the maximum (i.e., duplicating information already in the CV column under the four separate dates).
These CVs have all been calculated from one study's data (the Duffy data set, as denoted in column 1). Among-wetland estimates of variation are available from three other data sets. The "Hanson activity traps" data set has a single entry meeting the query criteria. The "a4" in the Wets column indicates that the CV estimate is from a comparison of four wetlands, and an "x" in the Yrs, Months, Veg, and AllSamps columns indicates that samples from an unspecified number of years, months, vegetation types, and replicates were pooled to represent each of the four wetlands, prior to calculating the CV. The CVmin column reports the smallest CV that any of the four wetlands had from the pooling of these other variables; CVmax reports the largest.
The "Euliss sweep net" data set has two entries that meet the selection criteria. One is a comparison of richness among 18 wetlands (a18 in the Wets column) and the other, a similar comparison among 16 wetlands (a16 in the Wets column). The difference is that each inter-wetland CV is from a different year. As indicated in the Years column, the 18-wetland comparison is from 1993 and the 16-wetland comparison is from 1992. The "x" in the Trans and AllSamps columns indicates that for each among-wetland comparison, data from an unspecified number of transects and replicates were pooled prior to calculating the CV.
Finally, the "Driver 1977" data set has two entries that meet the criteria. Both involve a comparison of 11 wetlands (a11 in the Wets column), in which data from an unspecified number of months and years had been previously aggregated (indicated by an "x" in the Months and Yrs columns). The difference between the two entries is apparent from viewing the WetTypes column: one CV estimate is from a comparison of 11 semipermanent (sp) wetlands; the other is from 13 temporary (t) wetlands. Also note, from the "Specific Taxon" column, that these richness estimates pertain only to the Chironomidae (midges). Taken as a whole, these CV estimates tentatively suggest that variability in midge species richness, as sampled by emergence traps, is greater among temporary wetlands (CV = 39.00) than among semipermanent wetlands (CV = 31.46).
Conceptually, the same approach as described in the above paragraphs can be used to scan the data base for estimates of other types of variability. For example:
- To review estimates of interannual variability, retrieve records that have a lower-case "a" in the Yrs column and proceed as described above.
- To review estimates of variation among zones within a wetland, retrieve all "a" records in the Zones column and do likewise.
- To compare two metrics (e.g., numindiv and biomass) with regard to their variability, retrieve all records for these as indicated in the Metrics column, and then match records according to other characteristics to make the comparison of CVs as fair as possible.
- To compare two major groups (e.g., invertebrates and plants) with regard to their variability, retrieve all records for these as indicated in the Group column, and then match records according to other characteristics to make the comparison as fair as possible.
Query #2
Does precision appear to increase the most when samples from a particular study are grouped by year, by season, by wetland, by zone, or some combination of these?
Approach:
First, go to the AllSamps column and for the particular study/data set and metric of interest, find all entries consisting of a number preceded by an "x". For example, for the data set on the Hanson activity traps and the metric "numiindiv," locate the entry "x320." This indicates that the associated CV estimate is based on 320 samples. The simultaneous presence of an "x" in the columns for Wets, Veg, Yrs, and Months indicates that the 320 samples constitute all the samples from varied wetlands, vegetation conditions, years, and months.
This CV (281) can then be used as a baseline against which to compare the CVs based on various subsets of the 320 samples. For example, when data are grouped by month (a5 in the Months column), the CV drops from 281 to 92. The value in the CVmin column of the same row indicates that under one vegetation condition in one wetland during one year, the variability among months was as low as 78. Likewise, the value in the CVmax column indicates that under one vegetation condition in one wetland during one year, the variability among months was as high as 223.
If instead of month, the data are grouped by year (a2 in the Yrs column), the CV drops more significantly from 281 to 80. The value in the CVmin column of the same row indicates that under one vegetation condition in one wetland during one month, the variability between years was as low as 119. Likewise, the value in the CVmax column indicates that under one vegetation condition in one wetland during one year, the variability between years was as high as 254.
A third option is to group the data by both year and month. This is indicated by a number (10, in this case) preceded by an upper-case "A" in both the Yrs and Months columns. The value in the CVmin column of the same row indicates that under one vegetation condition in one wetland, the variability among the ten possible combinations of year and month was as low as 61. Likewise, the value in the CVmax column indicates that under one vegetation condition in one (probably another) wetland, the variability among the ten year-month combinations was as high as 206.
Finally, note the few records where the AllSamps column has an entry consisting of a number preceded by an "x" (e.g., x320) and none of the other columns have an entry preceeded by an "x." The CVs for these records represent the variability among what are apparently true replicate samples, i.e., samples collected at the exact same location at the same time with the same equipment.
Appendix O.
Results of Asymptotic Richness Calculations Using Existing Prairie Data Sets
This table documents the results of the asymptotic richness calculations. Methods were described on page 15. Column headings are as follows:
Data Set
The source of the data; the experimental design associated with the analyzed data is detailed in Appendix L.
Equipment
Sampling equipment used by the collector of the original data.
Taxa
The category of organisms for which species accumulation rates were determined.
Samples
The type of samples (e.g., wetlands, quadrats, years) among which the taxa were accumulated.
Location
General location of the sampling; for BBS data:
stratum 37 = central prairie region
stratum 38 = western prairie
stratum 40 = eastern prairie
Total Taxa
Cumulative number of taxa in all samples of the analyzed data set.
Tot_NumSam
Total number of samples (wetlands, quadrats, years, etc.) among which species lists were combined and accumulated.
Num_for50...etc.
Number of samples (of the type specified) that are required to detect 50, 75, 90, 95, and 99% of the total taxa cumulatively present in the data set. The values given are the median for that threshold, based on the 100 random runs.
CurveTurn
A value of 0 indicates a curve terminating in a plateau (i.e., no slope). The larger the value, the farther from an asymptotic condition is the near-final part of the curve and the more likely that the sample was insufficient to estimate richness. Values were calculated as:
100 * (b-a) / (c-b),
where a = median for 90th percentile (Num_for90)
b = median for 95th percentile
c = median for 99th percentile.
CurveEnd
A value of 100 indicates that less than the full number of samples would have been required to sufficiently estimate richness; the smaller the value, the likelier is it that the sample size was more than sufficient, i.e., oversampling occurred. Values were calculated by dividing Num_for99 by Tot_NumSam and multiplying by 100.
Qualifiers
Supporting information that describes the exact type of data used in the analysis.
PooledVars
Variables that were subsumed in defining the samples (records) that were used for the analysis.
1. 1 Note that although the following species (which were not included) are usually believed to depend more on uplands than on wetlands, they were found in more than 5 (10%) of the North Dakota prairie areas classified as wetlands by Igl and Johnson (unpub. data, NPSC, Jamestown, ND): brown-headed cowbird, western meadowlark, eastern kingbird, morning dove, common grackle, horned lark, western kingbird, American robin, clay-colored sparrow, upland sandpiper, gray partridge, vesper sparrow, lark bunting.
2. 2 A "plot" is a set of 1 to 6 wetlands, generally of diverse water regime types and all located within a 4-mi2 area.
3. 3 Called "polygon" in the Euliss data set.
4. 4 Software for sorting the values is not provided. Users must either sort manually or import the file into a program such as dBase or Paradox.