A predictive model for assessment of stream health in the Darwin-Daly region of the Northern Territory using mostly genus level data. P.L. Dostine Report 42/2004D Natural Resource Management Conservation and Natural Resources PO Box 30 Palmerston NT 0831
27
Embed
A predictive model for assessment of stream health in the ...territorystories.nt.gov.au/bitstream/10070/213101/1/NT_genus_model.pdf · selected and sorted for macroinvertebrates until
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
A predictive model for assessment of stream health in the Darwin-Daly region of the Northern
Territory using mostly genus level data.
P.L. Dostine
Report 42/2004D Natural Resource Management
Conservation and Natural Resources PO Box 30 Palmerston NT 0831
CONTENTS Summary 3 Reference site selection procedures 3 Timing of sample collection 4 Laboratory analyses of macroinvertebrate samples 5 Model construction methods 6 Model assessment 13 Conclusions 15 References 15 Appendices 16
2
Summary This report details the methods and results of a project to construct a predictive model for the ecological assessment of stream and river condition in the Darwin-Daly region of the Northern Territory. The Darwin-Daly region comprises the catchments of the Mary, Adelaide, Finniss and Daly Rivers as well as several smaller rivers draining into Darwin harbour. The region includes two major population centres (Darwin and Katherine) and an area with high potential for agricultural development and catchment disturbance (Daly-Douglas area). The model was developed from samples of macroinvertebrates collected over several years in the early Dry season from channel edge habitat, using data at predominantly genus level (83% of taxa). Macroinvertebrate samples were preserved in the field and sorted in the laboratory using standard protocols. The sorting procedure employed a box-type sub-sampler and channelled sorting tray to randomly select at least 200 organisms. Specimens were identified to genus using regional taxonomic keys and/or taxonomic specialists. The raw data was culled by random sampling to provide a sample of 200. The final dataset comprised 117 reference individual sites and 198 taxa. All taxa were retained in the dataset. Classification of presence/absence data identified seven discrete groups. Overall these groups do not show strong geographic patterning, though some groups are geographically confined e.g. group 2 (15 sites) is mostly confined to the Mary River; group 3 (14 sites) is mostly confined to the lower Daly River; and group 7 (5 sites) is confined to the upper Adelaide/upper Finniss area. This lack of strong geographic patterning is not a surprising result given the lack of steep altitudinal and climatic gradients in the region. Twenty-five habitat variables were used in Multiple Discriminant Function Analysis to discriminate between classification groups. There were seven significant predictor variables: (i) distance from source; (ii) average habitat current velocity; (iii) standard deviation of elevation in adjacent DEM grid cells; (iv) the area of riparian rainforest within a radius of 100 m from the site; (v) stream order; (vi) alkalinity; and (vii) average stream width. The re-substitution error was 0.3868, and the cross-validation error was 0.5621. These and other model parameters are summarised in Table 1. Reference site selection procedures The stream reference dataset includes samples from sites collected during the Australian Rivers Assessment System (AUSRIVAS) program and the First National Assessment of River Health (FNARH) stages 1 and 2 of the Monitoring River Health program in the Northern Territory; from sampling conducted by an NHT funded project on water quality monitoring in the Mary River catchment; and on-going water quality monitoring program in catchments draining into Darwin Harbour. Samples were collected within five river basins (Table 2). Figure 1 shows the distribution of reference sites in the Darwin-Daly region. The model was generated using data from 117 sites. All sites were considered to be in good ecological condition following subjective assessment of the level of catchment disturbance due to land clearing, pastoralism, mining contamination and suburban development.
3
Region Darwin-Daly region, NT Taxonomic level Genus (83% all taxa) Model type Early Dry season, channel edge habitat Number of reference sites 117 Number of reference samples 117 Collection date of reference data 1996 18 1997 1 1998 19 1999 20 2000 47 2001 8 2002 4 Number of taxa 198 Number of reference site groups 7 Number of predictor variables 7 Distance from source (km) Standard deviation of elevation Average current velocity Alkalinity Stream width Area of riparian rainforest within 100 m radius from site Stream order Band width 0.27 Band thresholds A 1.14 B 0.85 C 0.57 D 0.29 Cross-validation error rate 0.5621 Resubstitution error rate 0.3868 Table 1. Summary of data input and model parameters for Darwin-Daly genus model.
River basin Reference sites Adelaide River 6 Darwin catchments 8 Finniss River 19 Mary River 34 Katherine-Daly River 50
Table 2. Distribution of reference sites. Timing of sample collection Samples were collected in the early dry season over seven years from 1996 to 2002. Most (>73%) were collected in the three year period from 1998 to 2000. Samples were collected from early May to mid-August, however, most (>70%) were collected over a six week period from late June to early August (Table 3). Over 80% of samples were collected by the end of July. Stream flows at this time are low and declining from the high and variable flows of the previous wet season.
4
Date 1996 1997 1998 1999 2000 2001 2002 Total % tot Cum
% tot 6-10 May 2 2 1.7 1.7 11-15 May 1 1 0.9 2.6 16-20 May 2.6 21-25 May 2 1 3 2.6 5.1 26-30 May 1 2 3 2.6 7.7 31 May-4 Jun 1 3 4 3.4 11.1 5-9 Jun 4 3 3 10 8.5 19.7 10-14 Jun 1 1 2 1.7 21.4 15-19 Jun 1 2 3 2.6 23.9 20-24 Jun 4 5 1 2 12 10.3 34.2 25-29 Jun 2 3 2 7 6.0 40.2 30 Jun-4 Jul 3 2 5 4.3 44.4 5-9 Jul 1 9 3 13 11.1 55.6 10-14 Jul 3 1 2 6 5.1 60.7 15-19 Jul 1 8 9 7.7 68.4 20-24 Jul 3 3 2.6 70.9 25-29 Jul 2 7 1 4 14 12.0 82.9 30 Jul-3 Aug 2 11 13 11.1 94.0 4-8 Aug 3 3 2.6 96.6 9-13 Aug 1 2 3 2.6 99.1 14-18 Aug 1 1 0.9 100 Total 18 1 19 20 47 8 4 117 Table 3. Number of samples collected in five day periods in each of years 1996-2002. Laboratory analyses of macroinvertebrate samples Laboratory procedures followed methods previously reported for NT AUSRIVAS model development (Dostine 2001) with the following differences. Samples were washed under tap water in a 500 μm sieve, rather than a 250 μm sieve. Sieved macroinvertebrate samples were sub-sampled using a box-type sub-sampler. Cells from the sub-sampler were randomly selected and sorted for macroinvertebrates until the cumulative total exceeded 200 animals and the last cell was completely sorted. Microcrustacea including copepods, ostracods and cladocerans were not included in the tally. A sample of individuals of large taxa present in the sample were collected and tallied separately as “extras” but were not included in analyses. For several samples, particularly in the Mary River NHT project (Schultz et al. 2002), the cumulative tally exceeded 400 animals. Data from these, and all other samples, were randomly sub-sampled to give a total of 200 animals using the following procedure. All individuals in each sample were arranged serially by taxa code number in the computer spreadsheet MS Excel. Some organisms which could not be identified to genus were deleted (generally some pupal Chironomidae, and some larval Coleoptera). The remainder were allocated a randomly generated number and sorted: the first 200 animals were regarded as the final sample for that site. Samples were identified as far as possible to the level of genus. A number of groups including chironomids and adult beetles were identified to species or species-type, however only genus level data were used in analyses. All chironomids were identified by P. Dostine; most beetles (except elmids) were identified by Dr Chris Watts (South Australian Museum); and most hydracarinids were identified by Dr Jane Growns. Elmid beetles were identified by Ms Alena Glaister (Monash University). All other material was identified by P. Dostine, or by Cyrus Edwards under close supervision of P. Dostine. There was a QA/QC program for samples from the Mary River catchment. This is not reported here.
5
Model construction methods The NT genus model was constructed by Ms Amara Barlow of the Cooperative Research Centre for Freshwater Ecology (CRCFE) at the University of Canberra under the supervision of Richard Norris. All model outputs reported here were derived by analyses conducted by the CRCFE. Data formatting
The raw data was checked for missing values and inappropriate variables. A total of 117 sites and 198 macroinvertebrate taxa were used in model construction. Total site taxa richness ranged from 22-61 taxa. Relative abundance data (i.e. number per 200 organisms) were transformed to presence/absence data. Macroinvertebrate taxa used in the model are listed in Appendix 1. Taxa codes are derived from the freshwater taxa code system developed by Dr John Dean.
Rare taxa
Rare taxa are generally removed from AUSRIVAS models. All taxa were retained for the NT genus level model.
Classification Site groups based on similarity of macroinvertebrate composition, were determined using cluster analysis in the PC-ORD multivariate analysis package, version 4.20 (McCune and Mefford, 1999). The Bray-Curtis distance measure and the agglomerative clustering technique, flexible Unweighted Pair-Group arithMetic Averaging (UPGMA), were used to form reference site groupings. The classifications were viewed as dendrograms allowing the fusion level, which divides sites into groups, to be selected. A beta value of -0.1 was used, as recommended for flexible UPGMA classifications (Belbin and McDonald, 1993) (Figure 2). The 117 reference sites were classified into seven groups. Overall, there is no strong geographic patterning in the distribution of site groups (Table 4). Group 1 (30 sites) includes sites from all 5 river basin areas but predominantly from the Daly and Finniss Rivers. Group 2 (15 sites) includes sites from 3 river basins but mostly from the Mary River. Group 3 (14 sites) includes sites mainly from the lower Daly River area. Group 4 (11 sites) includes few sites from the Daly, Mary and Darwin catchments. Group 5 (33 sites) is dominated by sites from the Mary River. Group 6 (9 sites) is a small group with 3 sites from the Daly River tributary, the Douglas River. Group 7 (5 sites) is a small group with 2 sites from the Finniss River and 3 sites from the Adelaide River. Site AD13 was classified as an outlier related to groups 5 and 6. It is included in group 6 in the table below. The sites used in the classification and their final group membership are listed in Appendix 2. The distribution of reference sites within each classification group are shown in Figure 1.
6
Figure 1. Location of Darwin-Daly region; location of reference sites within Darwin-Daly region; and location of classifications groups 1-7.
7
Figure 1 (cont). Location of Darwin-Daly region; location of reference sites within Darwin-Daly region; and location of classifications groups 1-7.
AD14 7 ----------| |--------------------| AD10 7 ----------------------------|--------| | FN23 7 ----------------------------| |----| FN09 7 -------------------------------------| Figure 2. Dendrogram of UPGMA classification of presence-absence macroinvertebrate data from 117 reference sites in Darwin-Daly region showing group numbers.
Table 4. Distribution of reference sites in each classification group by river basin. The mean number of taxa per classification group ranged from 34.8 to 42.8. Within classification groups there was substantial variability in the number of taxa (Table 5).
Group Mean number taxa Range of no. taxa per group
Table 5. Mean number of taxa and range of number of taxa per classification group.
Discriminant function analysis The reference site groups from the UPGMA classification were entered as a variable into the reference habitat data set and a stepwise Multiple Discriminant Function Analysis (MDFA) performed to select the predictor variables used in the AUSRIVAS model. The Stepwise MDFA was performed in the SAS statistical package version 6.12 (SAS Institute 1995). This procedure selected a subset of habitat variables that best discriminated among the groups of sites formed from the faunal classifications. The stepwise procedure included habitat variables one at a time, selecting at each step the variables that gave the best group discrimination. At each step of the analysis the significance of variables already included were checked and variables that were no longer significant were removed. The significance levels for variables to enter and be retained by the stepwise MDFA both were set at 0.05. The subsets of habitat variables from the Stepwise MDFA were tested in a MDFA to predict the probabilities of group membership for a reference site. Biased discriminations were avoided by using the cross-validation option that predicts group membership of each site separately. The subsets of habitat variables obtained from the stepwise MDFA are used as predictor variables for the AUSRIVAS model under construction. The predictor variables and the reference site invertebrate-classification form the foundation of AUSRIVAS, allowing predictions of the taxa expected at new “test” sites. Discriminant Function analyses were repeated using several versions of the habitat data to consider the effects of log transformation, the effects of merging classification groups (3 and 4; 5 and 6) and the effect of including or excluding the outlier site AD13. Outputs from each of these options are presented in Table 6.
5 7 groups, untransformed variables, AD13 included in group 5
0.3912 0.5607 0.29 -0.56 1.05 0.325
6 7 groups, log variables, AD13 included in group 5
0.3274 0.4787 0.32 -1.35 1.08 0.308
7 7 groups, log variables, AD13 included in group 6
0.3832 0.5323 0.29 -1.56 1.09 0.317
8 7 groups, untransformed variables, AD13 included in group 6
0.3868 0.5621 0.27 -1.33 1.09 0.333
Table 6. Re-substitution and cross-validation errors, band width and regression results for eight model options. Model options were assessed by considering re-substitution and cross-validation errors, band width, and the terms of the regression equation between observed and derived expected values. After consulting senior CRCFE staff model option 8 was selected primarily due to the relatively high R2 value and the relatively small band-width. The value of the misclassification error is not critical because AUSRIVAS uses all the probabilities of a site belonging to each group for site predictions, rather than the allocation to a single group as performed by the cross-validation procedure. Thus, sites with an affinity for two or more groups can be misclassified but still provide adequate predictions for a model. Predictor variables for the NT genus model are presented in Table 7. The list of habitat variables used in the MDFA is presented in Appendix 2. Means and standard deviations for predictor variables for each classification group are presented in Figure 3 and Table 8. The distribution of reference sites by river basin and mean site characteristics for each classification group are summarised in Table 9.
Predictor variable code Description and units
DFS Distance from source (kilometres)
STORDER Stream order
STREAMWIDTH Average stream width of the reach
ALKALINITY Alkalinity (Ca CO3 derived in the laboratory)
AVVELOCITY Average water velocity (metres/second)
SDVTOP Standard Deviation DEM Window of 5x5 cell window (derived from GIS coverage – standard deviation of elevations from 25 cells centred on collecting site derived from digital elevation model)
RIP100 Area of riparian rainforest within 100 m radius of site
12
Table 7. Predictor variables determined by MDFA for Darwin-Daly genus model.
WIDTH Std dev 18.9 4.9 24.1 17.4 7.3 4.3 10.3 Mean 70.0 90.2 248.2 92.1 36.2 144.4 74.0 ALKALINITY Std dev 113.9 148.7 71.2 131.2 52.6 98.6 50.0 Mean 0.007 0.016 0.047 0.038 0.074 0.150 0 000 AVVELOCITY Std dev 0.034 0.024 0.052 0.060 0.070 0.054 0.000 Mean 6.8 1.7 2.9 1.8 2.4 3.7 1.1 SDVTOP Std dev 7.1 0.8 2.0 1.1 1.9 4.6 0.9 Mean 0.0009 0.0016 0.0037 0.0006 0.0020 0.0040 0.0014 RIP100 Std dev 0.0015 0.0018 0.0018 0.0014 0.0024 0.0018 0.0019
Table 8. Means and standard deviations for seven predictor variables for each classification group.
Group No. Distribution Site characteristics
1 30 Daly (15), Finniss (10) High SDVTOP 2 15 Mary (9) AVGWID mean 9.5 m 3 14 Daly (12) High DSTSRC, STRORD, RIP100, AVGWID, ALKNTY 4 11 Daly (5), Mary (3), Darwin (3) Low RIP100 5 33 Daly (9), Mary (20) Low ALKNTY, medium AVGVEL 6 9 Daly (5) (Douglas (3)) AVGWID mean 9.4 m, high AVGVEL, medium
SDVTOP, high RIP100 7 5 Adelaide (3), Finniss (2) STRORD mean 3.8, low SDVTOP, AVGVEL zero
Table 9. Number of reference sites in each classification group, distribution by river basin and characteristic features of each classification group. Model assessment The band distribution O/E values of reference sites for the model is 72.7% (85) in band A, 17.9% (21) in band X and 7.7% (9) in band B. When setting the O/E taxa bands only the O/E taxa range between the 10th and 90th percentile from the population of reference sites used to create the model is considered as equivalent to reference condition. A disproportionately high number of sites were classified as X (above reference condition). Two sites (DA40 and FN14) were designated as being out of the range of the model. The assessment of reference–test sites identified 1 site (DA40) as outside the experience of the model, 11 sites in band A and 3 sites in band X (Appendix 5). Three sites have been sampled on either 3 or 4 occasions between 1995 and 1999. OE50 values for these sites are shown in Table 10. Site 1995 1996 1998 1999DA03 1.13 1.08 1.08DA04 1.18 1.12 1.14 1.02DA11 1.14 1.05 1.01 1.07Table 10. OE50 values for three sites sampled on at least three occasions.
13
1 2 3 4 5 6 70
100
200
300
400
500
600
1 2 3 4 5 6 70
2
4
6
8
1 2 3 4 5 6 70
10
20
30
40
50
60
70
1 2 3 4 5 6 70
50
100
150
200
250
300
350
1 2 3 4 5 6 70.00
0.05
0.10
0.15
0.20
0.25
1 2 3 4 5 6 70
2
4
6
8
10
12
14
16
1 2 3 4 5 6 70.000
0.001
0.002
0.003
0.004
0.005
0.006
0.007
Group number
Group number
Group number
Group number
Group number
Group number
Group number
Dis
tanc
e fro
m s
ourc
e (k
ms)
Stre
am o
rder
Aver
age
wid
th (m
)Al
kalin
ity
Aver
age
velo
city
(m/s
)St
d de
v el
evat
ion
Are
a rip
aria
n ra
info
rest
(a)
(b)
(c)
(d)
(e)
(f)
(g)
Figure 3. Mean values and standard deviations for seven predictor variables by classification group. (a) Distance from source; (b) stream order; (c) average stream width; (d) alkalinity; (e) average velocity; (f) standard deviation elevation; (g) area of riparian rainforest with 100 m radius.
14
Conclusions
A predictive model based on macroinvertebrate composition has been developed for assessment of stream and river condition in the Darwin-Daly region of the Northern Territory. Data used for model construction was predominantly at the level of genus. Previous models developed by the NT AUSRIVAS program have used family level data and sites across the Top End of the Northern Territory. Genus level data would provide better discrimination between site groups and potentially better models, though these issues are yet to be addressed.
The model is based on data from 117 sites from channel edge habitats from the early Dry season. Macroinvertebrate data for each sample were standardised through random deletion (or in some cases additional sampling) to provide a total of 200 specimens. Data from large organisms not included in the random sorting process (classified as “extras”) were not included in analyses. Most samples were collected during a six-week period from late June to early August.
Classification of presence-absence data identified seven groups of sites. Site groups do not have strong geographic patterning, and are distributed across catchments within the model area. Most predictor variables refer to position within the catchment and other landscape characteristics. Fine-scale predictor variables were generally unavailable for assessment, or were considered to be subject due to unacceptable biases.
The model yielded satisfactory results for “reference-test” sites. Eleven of fifteen reference-test sites (73%) were correctly identified as in reference condition. Three of the remaining four sites were marginally above the band threshold and were classified as above reference condition (band X). The model has been applied to derive a relationship between catchment disturbance (%catchment cleared) (Townsend et al. 2004) and had been applied in the routine annual assessment of stream health in the catchments of the Darwin harbour area.
References
Belbin, L. and McDonald, C. (1993). Comparing three classification strategies for use in ecology. Journal of Vegetation Science, 4: 341-348.
Dostine, P. (2001). Australia-wide assessment of river health. Northern Territory program. Final report. Report to Environment Australia.
McCune, B. and Mefford, M. (1999). Multivariate Analysis of Ecological Data Version 4.20. Mjm Software, Gleneden Beach, Oregon, USA.
Schultz, T.J., Townsend, S.A., Edwards, C.A. and Dostine, P.L. (2002). Water quality monitoring in the Mary River catchment. Tech. Report 42/2002. Department of Infrastructure, Planning and Environment, Darwin, NT.
Townsend, S.A., Dostine, P.L. Dixon, I., Karfs, R. and Douglas, M.M. (2004). The response of stream macro-invertebrate communities to catchment clearing and riparian condition in the Darwin region (tropical Australia). Proceedings of the 4th Australian Stream Management Conference (in press).
15
Appendices
Appendix 1. List of taxa used in development of NT genus-level model.
Appendix 2. Sites used in model construction and their group membership as determined by UPGMA classification of macroinvertebrate presence-absence data.
Appendix 3. Predictive variables used in Multiple Discriminant Function Analysis.
Code Variable description LATUDE Latitude LNGUDE Longitude DAYNUM Day number from 1 Jan DSTSRC Distance from source CATARE Catchment area LOGCAT Log catchment area STRORD Stream order AVGWID Average stream width ALKNTY Alkalinity AVGVEL Average velocity SDVTOP Standard deviation of elevation RNGTOP Range of elevation ELVATN Elevation RIP100 Area of riparian rainforest in 100m radius of site SPR100 Area of spring rainforest in 100m radius of site TOT100 Total rainforest in 100m radius of site RIP250 Riparian rainforest in 250m radius of site SPR250 Spring rainforest in 250m radius of site TOT250 Total rainforest in 250m radius of site RIP500 Riparian rainforest in 500m radius of site SPR500 Spring rainforest in 500m radius of site TOT500 Total rainforest of 500m radius of site RIP1000 Riparian rainforest in 1000m radius of site SPR1000 Spring rainforest in 1000m radius of site TOT1000 Total rainforest of 1000m radius of site
23
Appendix 4. Outputs from NT Darwin-Daly genus model for reference site data from CRCFE analyses. NTE50 number of taxa predicted to occur with 50% probability; NTC50 number of taxa predicted to occur with 50% probability actually collected; OE50 ratio of observed/expected.
Site NTE50 NTC50 OE50 Band DW43 19.6 14 0.71 B MA33 20.5 15 0.73 B DA39 18.7 14 0.75 B DA17 20.6 16 0.78 B MA15 23.4 19 0.81 B DA14 24.1 20 0.83 B DW50 20.5 17 0.83 B MA27 21.5 18 0.84 B DA28 21.5 18 0.84 B MA16 20.9 18 0.86 A FN11 22.1 19 0.86 A MA08 22.0 19 0.86 A DA09 26.4 23 0.87 A MA14 20.7 18 0.87 A FN04 21.8 19 0.87 A MA20 21.3 19 0.89 A DA38 21.2 19 0.90 A MA13 22.2 20 0.90 A DA40 21.0 19 0.90 Out of range DA45 20.9 19 0.91 A AD06 19.8 18 0.91 A DA29 23.0 21 0.91 A DW40 19.7 18 0.91 A FN24 19.5 18 0.92 A DA42 21.7 20 0.92 A FN26 22.7 21 0.93 A DW48 22.5 21 0.93 A MA25 20.3 19 0.94 A MC03 19.2 18 0.94 A FN14 23.4 22 0.94 Out of range MA32 22.3 21 0.94 A FN06 21.2 20 0.95 A - Attention DA01 26.4 25 0.95 A MA07 22.2 21 0.95 A MA19 21.1 20 0.95 A DA13 22.1 21 0.95 A MA26 21.9 21 0.96 A FN02 19.7 19 0.97 A DA37 22.6 22 0.97 A DA60 22.6 22 0.97 A DA30 22.5 22 0.98 A DA54 26.4 26 0.98 A DW49 19.2 19 0.99 A AD14 19.2 19 0.99 A AD11 21.1 21 0.99 A DW51 24.1 24 1.00 A MA34 22.1 22 1.00 A DA41 21.0 21 1.00 A FN22 22.0 22 1.00 A
24
MC04 21.0 21 1.00 A MA28 18.9 19 1.00 A MA35 20.9 21 1.00 A DA21 20.9 21 1.01 A DW37 20.8 21 1.01 A AD13 18.8 19 1.01 A DA46 20.8 21 1.01 A DW39 18.8 19 1.01 A MC06 20.8 21 1.01 A DA05 23.7 24 1.01 A MA09 22.7 23 1.01 A AD10 20.7 21 1.01 A DA53 22.6 23 1.02 A DW45 18.6 19 1.02 A DW42 18.6 19 1.02 A DA58 26.4 27 1.02 A DA59 26.4 27 1.02 A DA04 19.6 20 1.02 A DA52 20.5 21 1.02 A DA32 26.3 27 1.03 A FN12 21.4 22 1.03 A DA50 21.4 22 1.03 A DA35 22.3 23 1.03 A MA24 22.3 23 1.03 A MA10 21.2 22 1.04 A DA47 19.2 20 1.04 A MA23 22.0 23 1.04 A FN23 21.9 23 1.05 A MA11 20.7 22 1.06 A DA49 22.6 24 1.06 A MA30 20.7 22 1.06 A DA11 21.5 23 1.07 A FN01 23.2 25 1.08 A DW36 21.3 23 1.08 A DA03 22.2 24 1.08 A MA17 21.1 23 1.09 A DA15 21.0 23 1.09 A DA22 21.0 23 1.09 A DA07 26.4 29 1.10 A FN13 20.0 22 1.10 A MA12 20.9 23 1.10 A DA57 20.9 23 1.10 A MA01 21.7 24 1.11 A DA43 20.7 23 1.11 A FN09 19.8 22 1.11 A MA02 21.3 24 1.13 A DA20 20.4 23 1.13 A DA24 20.9 24 1.15 X MA31 20.8 24 1.16 X MA18 19.8 23 1.16 X DA12 22.3 26 1.17 X MA21 22.2 26 1.17 X DA08 26.4 31 1.17 X – AttentionDA31 21.8 26 1.19 X
25
DA51 22.6 27 1.19 X DA44 22.5 27 1.20 X MA29 21.3 26 1.22 X DA56 21.0 26 1.24 X FN07 19.9 25 1.26 X DA48 21.5 27 1.26 X DA10 19.8 25 1.26 X DA55 21.3 27 1.27 X MA22 22.9 29 1.27 X AD02 21.3 27 1.27 X DW31 21.0 27 1.28 X DA27 26.4 34 1.29 X FN03 20.1 27 1.34 X DA18 21.3 29 1.36 X
26
Appendix 5. Outputs for NT Darwin-Daly genus model for reference-test sites.
Site NTE50 NTC50 OE50 Band DA01 26.4 24 0.91 A DA40 21.8 20 0.92 Out of rangeDA05 23.7 22 0.93 A DA21 21.1 20 0.95 A DA22 21.1 21 1.00 A DA11 (98) 21.8 22 1.01 A DA11 (96) 21.0 22 1.05 A AD02 21.7 23 1.06 A DA03 (99) 21.3 23 1.08 A DA04 (96) 19.6 22 1.12 A DA03 (96) 22.2 25 1.13 A DA04 (98) 18.5 21 1.14 A DA11 (95) 21.0 24 1.14 X FN04 20.8 24 1.15 X DA04 (95) 20.4 24 1.18 X