SEARCH

SEARCH BY CITATION

Keywords:

  • agricultural impacts;
  • environmental information;
  • New Zealand;
  • riparian;
  • stream and river health;
  • stream restoration;
  • water quality

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References
  • 1
    Little is known of farmer responses to environmental education, and where practices aimed at improving stream health are adopted, comparisons are generally limited to between-stream comparisons rather than monitoring changes over time. A 2-year study of farmer responses to the provision of information and associated assessment of stream health following voluntary adoption of measures to improve stream health was carried out on 30 deer farms in southern New Zealand. Farmers were allocated to groups that did (n = 14) or did not (n = 11) receive information about waterway management. An additional five farmers were not involved in the information provision study. Stream health indices from a stream on each farm (n = 30 streams) were assessed in March 2001, 2002 and 2003, before and after the voluntary adoption of best management practices (BMPs) aimed at improving stream health (n = 7 streams).
  • 2
    While some stream health indices responded as expected if BMP information provision influenced stream management, no response was statistically significant. Four information farmers, one no information farmer and two not involved in the information provision study adopted BMPs on streams, including permanent fencing to exclude stock from streams and ongoing management changes such as limiting grazing intensity in deer paddocks with waterways. After BMP adoption in these seven streams, environmental conditions changed in a consistent positive manner, whereas no consistent pattern was observed in streams where BMPs were not adopted. Invertebrate community indices also improved where BMPs were adopted.
  • 3
    Synthesis and applications. This ‘BACI’-type study evaluated effects of provision of information and voluntary adoption of a range of fairly minor BMPs for stream health. No significant effects were associated with the ‘information’–‘no information’ comparison, but in streams where BMPs were voluntarily adopted, consistent improvements in stream health indices occurred within 2 years. These differences were detected despite our inability to properly control reference stream conditions and using relatively simple field methods. Such positive outcomes soon after implementation of BMPs are encouraging, although longer-term appraisals of BMP adoption are required to assess whether the positive changes observed here would generate significant improvements in stream health.

Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

Acknowledging the role of human behaviour is critical for progress in conservation (Rhoads et al. 1999; Wang, Lassoie & Curtis 2006). While environmental education is widely used and assessments of environmental education are becoming more common, many are focused narrowly on participant attitudes, knowledge and awareness. Studies examining management outcomes that follow information provision are rare and few, if any, links have been made between information provision and ensuing environmental benefits.

In New Zealand, where lowland surface water quality is often degraded because of agricultural practices, environmental information is used to promote best management practices (BMPs) to improve stream health (Larned et al. 2004). BMPs are intended to mitigate a wide variety of agricultural impacts on streams (Withers & Jarvis 1998). While there is some evidence that provision of information may increase voluntary adoption of BMPs (Rhodes 2004), little is known of the in-stream benefits. The benefits of improved riparian management are generally accepted (Perry et al. 1999; Collier, Fowles & Hogg 2000) but evaluations focus usually upon protection of relatively long stream lengths (e.g. Quinn et al. 1992; Howard-Williams & Pickmere 1994) or entire catchments (Williamson, Smith & Cooper 1996). These large-scale management approaches may not represent BMPs that farmers adopt voluntarily (Gale et al. 1993), and assessments of other kinds of management modifications are rare (Quinn et al. 1992; Williamson, Smith & Quinn 1992; Sovell et al. 2000; Weigel et al. 2000).

Existing studies frequently lack critical aspects of rigorous experimental design. Many substitute space for time (Quinn et al. 1997), comparing protected and unprotected reaches, matched for other variables (e.g. Storey & Cowley 1997; Scarsbrook & Halliday 1999; Rhodes 2004). Such studies may be confounded, as observed differences could be inherent to the reaches compared rather than due to the effects of management. Two New Zealand studies (Howard-Williams & Pickmere 1994; Williamson et al. 1996) and one in North America (Cook et al. 1996) have shown improvements over time in some water chemistry variables after implementation of diffuse pollution control measures, but none included a control.

We describe a two-part study with the following aims: first, to establish if changes in stream health over time are different in waterways managed by farmers who were provided with information pertaining to BMP of agricultural streams relative to those who were provided with no information; and secondly, using the same streams, to determine whether environmental responses differ in streams where BMPs are voluntarily adopted compared to streams where no intentional BMPs are instigated. We address the design shortcomings of previous investigations by using a semirandomized ‘before–after, control–intervention’ (‘BACI’; Green 1979) approach to determine relationships between information provision, the voluntary adoption of BMPs and stream health.

Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

study sites and information provision

A single permanent stream in the deer unit (paddocks fenced for deer grazing) of each of 30 farms in the province of Southland, South Island of New Zealand, was assessed in March 2001, 2002 and 2003 (early autumn). In each case, a single 20-m riffle site at the bottom of the most downstream paddock was studied. If streams were discoloured by rain events, sampling was delayed until clear flow resumed.

Twenty-five of the 30 streams were managed by farmers who were included in the concurrent assessment of the behavioural consequences of information provision on BMPs relating to stream management (14 farmers received information, 11 did not, between March and September 2001). These streams were used for the information–stream health investigation. Another three streams were on properties taken over by new managers during the study period, and two more were on properties never involved in the behavioural study, giving 30 streams for the study of BMP adoption on stream health. Management responses by farmers to the provision (or lack) of information were voluntary hence the ‘BMP-adopted’ stream group for this analysis comprised an unpredictable subset of streams.

Of the 25 farmers whose streams were used for the information–stream health investigation, 20 were assigned randomly to either information or no information groups, one insisted on receiving information and four did not want it. Stratified random allocation was used to ensure a similar range of stream sizes and conditions were present within each group. Farmers in the information group attended a stream management field day (or received a written summary), received pamphlets, and a personal visit from a Land Management Extension Officer from Southland Regional Council in the 4 months following the first field season (March 2001). Farmers’ receipt of information extraneous to that provided in the study was monitored, with both groups reporting similar levels of exposure during the 2 years of study (Rhodes 2004).

field and laboratory methods

Stream water nutrient concentrations were measured from duplicate water samples that were vacuum-filtered through pre-ashed and weighed glass microfibre filters (Whatman GF/F), transported on ice and frozen. Thawed samples were auto-analysed colorimetrically for nitrate/nitrite–nitrogen (hereafter referred to as NOx-N), ammoniacal nitrogen and dissolved reactive phosphorus (NH4-N and DRP, respectively) using standard methods (APHA 1998). Filter papers and accumulated sediments were frozen before drying to constant weight at 65 °C for calculation of total suspended solids (TSS).

To assess physical condition, visual estimates of substrate and bank cover were obtained from the study reach and the bottom 20 m of the next two paddocks upstream. The percentage of substrate fines (FINES –particles = 2 mm; Storey & Cowley 1997) was obtained by averaging visual estimates from the three 20-m reaches. The percentage of bank that was covered by bare soil or very sparse vegetation (assumed to be of little use for preventing surface run-off) was estimated visually along a 5-m wide strip on each bank (BARE).

A habitat score for each study site was obtained from the lower 10 m of the downstream 20-m study reach using the protocol described in the New Zealand Stream Health Monitoring and Assessment Kit (SHMAK) manual (Biggs, Kilroy & Mulcock 1998). The overall score was derived from six parameters (excluding the recommended water velocity and point measures of temperature because these were beyond farmers’ control): conductivity (TDScan 3, Eutech Instruments, Nijkerk, Holland); pH (pHep3 meter: Hanna Instruments, Rhode Island, USA); clarity (using a 1-m clarity tube); extent of loose deposited material in wetted area (visual estimate); bank cover (visual estimate of 10 × 5-m strip adjacent to each side of waterway) and substrate composition (Table 1). For the latter, substrate materials were characterized using the Wolman pebble count (Wolman 1954; Biggs et al. 1998). Points were awarded for each habitat parameter and a total SHMAK HABITAT score was calculated in the range –43 (poor stream health) to 80 (good stream health) (Table 1).

Table 1.  Scores awarded for values of parameters used to calculate SHMAK habitat score. The HABITAT score is the sum of the six individual contributing scores from each parameter
ParameterCategories and corresponding scores (in italic type)
pH≤ 5·45·5–6·46·5–7·98·0–9·4≥ 9·5
55105– 5
Conductivity (µS)≤ 5050–149150–249250–399≥ 400
20161061
Clarity (cm)Clear to end70–10055–6935–54≤ 35
108531
Stream bed composition (%)Bedrock > 25 cmCobbles 12–25 cmCobbles 6–12 cmCobbles 0·2–6 cmGravel
–101020100
[Contributing score = sum of all (indv. score,* %)]SandMud/siltManmadeWood debrisPlants
–10–20–2000
Extent of loose deposited material on stream bedNoneFine (less than 1 mm thick), mainly in edge areasModerate (≤ 3 mm), edge areas and elsewhereModerate to thick (≥ 3 mm), patchy, most of bedThick (> 5 mm), on most horizontal surfaces
(estimate scores between categories)1050–5–10
Bank vegetation (% of bank)Native treesWetland vegetationTall tussock grassland, not improvedIntroduced trees (willow, poplar) treesOther introduced (e.g. conifers)
[Contributing score = sum of all (indv. score* %) for each bank]1010885
ScrubRock, gravels improvedShort tussock grassland and weedsPasture grassesBare ground
5–53–10–10–10

Benthic macroinvertebrates were collected and identified at the downstream study reach of each stream using both the SHMAK in-field live-sorting identification procedure (low taxonomic resolution) for evaluation of stream health (Biggs et al. 1998), and a standard field sample of benthic material with later laboratory identification of invertebrates (higher taxonomic resolution). For the in-field SHMAK analysis, 10 samples of substrate (e.g. rock, macrophyte or mud) were collected randomly from the lower 10 m of each study reach, sorted and invertebrates were identified to the taxonomic levels prescribed by Biggs et al. (1998). The SHMAK invertebrate score (hereafter referred to as INV) awards points to different taxa based on their tolerance of pollution, with the overall score weighted for the abundance of each taxon in the sample. For laboratory analysis, we collected a standard 500 µm kicknet sample of benthic material for identification to higher taxonomic resolution, attaining the minimum level recommended by the New Zealand Macroinvertebrate Working Group (Stark et al. 2001) using the keys of Winterbourn, Gregson & Dolphin (2000) and Chapman & Lewis (1976). From this we calculated the Macroinvertebrate Community Index (MCI) and Quantitative MCI (QMCI) (Stark 1985). Although specific to New Zealand fauna, the MCI and QMCI are similar to other internationally used indices weighted for pollution tolerance [e.g. British Biological Monitoring Working Party score system (Hawkes 1998)]. Although the thresholds for these invertebrate indices are designed for application to stony streams (Stark 1985), the indices are expected to reveal impacts associated with land-use change in silty/sandy streams (Boothroyd & Stark 2000).

accounting for background variation

On each sampling occasion, farmers were surveyed to identify ‘BMP adopted’ streams where management changes had been made to improve stream health, as well as to record other routine variation in farm activities around and upstream of the study site that might account for some background variance in measured parameters of stream health. From this survey, we created five covariables representing: grazing in the catchment (presence or absence at time of sampling, as well as longer term); grazing in the study paddock (all grazing variables used stock unit conversions from Cornforth & Sinclair (1984)); ploughing and fertilizing. Variables were attenuated (by dividing by square root of number of weeks since action) to account for the non-linear decrease of in-stream impacts over time from grazing, plowing and fertilizing [for example, the rate of pasture recovery from grazing is non-linear (Brougham 1955)]. We also included ‘stream discharge at the time of sampling’ as a covariate.

statistical methods

Analysis was carried out in two parts. The first identified differences between changes in stream health between March 2001 and March 2003 in streams of farmers who had received information (‘information’ streams) and those who had not (‘no information’ streams). The second again identified differences between changes in stream health over the same time period, but compared streams where BMPs were adopted (‘BMP-adopted’ streams) to those where management was unchanged (‘reference’ streams). The same analysis approach was used for both comparisons.

Data from each stream in 2001 and 2003 were ordinated using principal components analysis with all variables centred and standardized, with regard to eight variables to reflect overall stream health: NOx-N, DRP, NH4-N, TSS, BARE, FINES, HABITAT and INV. Ordinations were performed using canoco for Windows version 4 (ter Braak & Smilauer 1998). Ordinations were performed with and without covariates; results are presented from ordinations which best distinguished ‘information’ from ‘no information’ sites, or BMP-adopted from reference (no BMPs adopted) sites. Where necessary, variables and covariates were transformed to improve normality. Differences in the overall distribution of change demonstrated by information and no information streams, and BMP-adopted and reference streams were tested using Watson's non-parametric test for equality of circular distributions using angular measurements in degrees (Zar 1999).

We used repeated-measures analysis of variance (anova) in spss release 11 (SPSS Inc. 2001) to investigate the univariate responses of the SHMAK habitat and invertebrate indices, as well as the more rigorously obtained MCI and QMCI. In these analyses, real differences between the trends over time shown by the information and no information groups, or the BMP-adopted and reference groups, are identified by a significant ‘group × year’ interaction term. We also checked for the influence of the eight background management variables using repeated-measures analysis of covariance (ancova), and subsequently included any significantly influential variable as a covariate in each analysis. All tests presented meet the requirements of homogeneity of variance (Box's M) and normality of residuals (assessed visually). Variables were transformed as previously to achieve normality.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

initial stream conditions

Initial water chemistry, habitat and invertebrate metrics varied across the 30 streams sampled, and most streams showed at least some impacts associated with pastoral farming (Table 2). Banks of many streams had little ground cover (high BARE) and much of the substrate was silt or sand (high FINES). The mean and median values of NOx-N in the streams were well above the maxima recommended (ANZECC 2000; Boothroyd & Stark 2000), and NH4-N was higher than the recommended level in many streams (Table 2). The average HABITAT score of 6 was below halfway on the possible range of –43 to 80. Invertebrate communities in all streams scored below the QMCI threshold value of 4 at different times, indicating ‘probable moderate pollution’, and 13 streams had scores suggesting ‘probable severe pollution’ (Boothroyd & Stark 2000; Table 2).

Table 2.  Mean, standard error and range for each parameter obtained from 30 study streams in March 2001. Length refers to mean distance between upper and lower sampling sites. Mean discharge measured at summer baseflow over 3 years
ParameterMeanSEMedianRangeCritical value
  • a

    Guideline for maximum recommended level for protection of aquatic health and ecosystems, periphyton limitation and preservation of water quality and aesthetics (ANZECC 2000).

  • b

    Threshold value below which indicates ‘probable severe pollution’ for stony streams (Boothroyd & Stark 2000).

  • c

    Threshold value below which indicates ‘probable moderate pollution’ for stony streams (Boothroyd & Stark 2000).

Length (m)1037129825300–2950
Discharge (L s−1)  11·8  3·2  6·250·2–70·4
Width (m)   2·4  0·3  2·20·6–8·1
Depth (cm)  11·9  1·1 11·52·9–27·9
Gradient (%)   1·2  0·1  10·3–2·5
TSS (mg L−1)  23·3  6·6  8·50·4–177·1
NOx-N (µg L−1) 515·7146·3178·94·9–3702·9100a
DRP (µg L−1)  17·1  3·8 11·92·4–110·2 30a
NH4-N (µg L−1) 508·1198·0 35·11·7–4478·0 30a
Bare (%)  52·4  4·6 53·35·0–100·0
Fines (%)  66·6  5·0 71·75·0–100·0
HABITAT score   6·0  3·3  6·0–28·0–50·7
INV score   3·37  0·19  3·23–1·56–5·72
No. of taxa  12·1  0·6 11·56·0–19·0
MCI score  78·2  2·1 80·053·3–98·880b, 99c
QMCI score   2·88  0·18 2·911·18–5·134b, 4·99c

changes in management initiated during study period

During the 2-year study period, BMPs were adopted at seven of the 30 streams (Table 3); four in the informed farmer group, one by an ‘uninformed’ farmer and two by farmers not involved in the information study. Adopted practices ranged from structural developments such as permanent fencing to exclude stock, to ongoing non-structural changes in management, such as reduction of grazing pressure under wet conditions. Four of the farmers adopted BMPs soon after the March 2001 stream health assessment, two made changes 1 or 2 months before September 2002, and one changed management only 1 month before March 2003 (Table 3). Because of this variation, data for BMP-adopted streams in March 2002 includes three streams at which management had not yet been changed.

Table 3.  Details and dates of implementation of BMPs that were voluntarily adopted on seven streams (numbered) to improve water quality and/or overall stream health
StreamChange/s in managementPeriod implemented
  • a

    Streams at which the farmer was in ‘information’ group of the behavioural intervention study (3, 8, 13, 17).

  • b

    Farmer not involved in information provision study (5, 30).

  • c

    Farmer in ‘no information’ group (21).

  • d Note that for March univariate analyses, no change had occurred for March 2002 time-point in these streams.

3aBuffer strip of at least 5 m left clear of fertilizer application Deer removed from waterway paddocks in wet conditionsPre-September 2001
13aDeer grazed for less time in waterway paddocks under wet conditionsPre-September 2001
30bTemporary electric fencing used to exclude deer from stream in paddocks above study paddocks (at first done onlywhen deer were grazing winter crops, but increasingly at other times)Pre-September 2001
21cUpper 150 m of stream from spring source permanently fenced to exclude stock (note that sample site is a further 235 m downstream in grazed area; buffer width 25 m true right, 7 m true left)3 months prior to September 2001
17aWaterway in study paddock (250 m) permanently fenced (buffer width average 10 m stream to fence; 10 m section at top of study paddock open for stock water but no grazing during study) dam at top of catchment fenced into smaller paddock for greater grazing control.5 months prior to September 2002d
8aWaterway in study paddock (200 m) permanently fenced into lane (stream-fence distance 13 m true right, 6 m on true left; deer grazed occasionally within fenced area, but less intensively)4 months prior to September 2002d
Small bog (0·4 Ha) upstream fenced off and drained 
New culvert installed 
5bWater trough installed in study paddock1 month prior to March 2003d

comparison of ‘information’ and ‘no information’ streams

Ordination of the eight parameters of stream health across the 25 streams where landowners were involved in the behavioural study showed no consistency in change when grouped by ‘information’ or by ‘no information’ (Fig. 1), and no difference in the distribution of changes in these two groups (Watson's inline image = 0·0746, P > 0·5).

image

Figure 1. March 2001–03: PCA ordination of 11 streams on properties of farmers in ‘no information’ group (a) and 14 farmers in ‘information’ group (b) with regard to eight parameters of stream health in (c). All plots obtained from the same ordination. Numbers on arrows in (a) and (b) identify individual streams. Arrowheads indicate position in March 2003 and base position in March 2001. Abbreviations: SHMAK habitat (HAB) and invertebrate scores (INV), nitrate/nitrite-nitrogen (NOx-N), ammoniacal nitrogen (NH4-N), dissolved reactive phosphorus (DRP), total suspended solids (TSS), substrate fines (FINES) and proportion of bare stream bank (BARE). Explained variance: axis 1 = 40·5%, axis 2 = 17·0%.

Download figure to PowerPoint

Trends in INV scores showed a general decrease over time in streams on ‘no information’ farms while ‘information’ stream scores were comparatively stable (Fig. 2a), but there was no significant difference between changes in the two groups (repeated-measures anova interaction term F2,46 = 1·31, P = 0·280). QMCI and MCI scores followed similar and statistically non-significant patterns (F2,46 = 0·43, P = 0·654 for QMCI and F2,46 = 0·087, P = 0·426 for MCI; Fig. 2b,c). HABITAT scores increased more between March 2001 and March 2003 among ‘information group’ than ‘no information’ streams (Fig. 2d), but this difference was again not significant (F2,46 = 0·61, P = 0·549), especially when the background variables DISCHARGE and SDSP (stock days in study paddock) were taken into account (combined covariate effect F2,44 = 7·85, P = 0·001; resulting information effect F2,44 = 0·10, P = 0·907; no farm management variables were significant when included as covariates in any other information analyses).

image

Figure 2. Means (± SE) of stream health indices from 11 streams on properties of farmers not receiving study information (inline image) and 14 streams on properties of informed farmers (x), March 2001–March 2003. (a) INV scores from SHMAK assessment; (b) QMCI scores; (c) MCI scores; (d) HABITAT scores from SHMAK assessment.

Download figure to PowerPoint

comparison of bmp-adopted and reference (no bmp-adopted) streams

There was a clear difference in the direction of changes observed over the two-year sampling period in the seven BMP-adopted streams compared to the reference group of 23, when all streams were ordinated with respect to the same eight stream health measures (Fig. 3a–c). Including background management covariates in the ordination did not improve this distinction.

image

Figure 3. (a–c) March 2001–03: PCA ordination of seven ‘BMP adopted’ streams (a) and 23 reference streams (b) with regard to the eight stream health parameters in (c). All three plots obtained from the same ordination, using similar data as for Fig. 1 but regrouped into ‘BMP adopted’ and reference streams and including all 30 sites. Numbers on each arrow identify individual streams. Arrowheads indicate position in March 2003, base indicates position in March 2001. See Fig. 1 for abbreviations. Explained variance: axis 1 = 42·2%, axis 2 = 15·0%.

Download figure to PowerPoint

Stream health at BMP-adopted sites changed in a generally consistent manner, shifting towards upper and left-hand quadrants in ordination space between March 2001 and 2003 (Fig. 3a). While trends for some reference streams were similar to those shown by BMP-adopted streams, there was no consistent pattern among these streams (Fig. 3b). Overall, the distribution of change in the two groups was significantly different (Watson's inline image = 0·3012, P < 0·005).

The dominant trend in BMP-adopted sites between years was an increase in HABITAT and INV scores, and a smaller decrease in FINES and BARE (compare Fig. 3a,c). Changes in nutrients and TSS did not appear to be related to the overall trend in treatment streams.

Univariate analyses showed that mean INV scores improved significantly in BMP-adopted streams over the 2-year sampling period, while those in reference streams declined (F2,56 = 9·747, P = 0·001; Fig. 4a). No management covariates altered this relationship significantly (minimum covariate P = 0·207). Because BMP-adopted and reference groups showed significantly different INV scores in March 2001 (T1,28 = 2·286; P = 0·009; note that no significant differences were found for any other ‘pretreatment’ comparisons), the analysis was repeated to compare the seven BMP-adopted streams with seven reference streams paired as closely as possible for initial INV scores (Fig. 4b). A significant positive treatment effect remained (F2,24 = 1·13, P = 0·007).

image

Figure 4. Means (± SE) of indices of stream health in seven ‘BMP adopted’ streams (inline image, solid line) and 23 (inline image, dashed line) reference streams over time. (a) INV score from SHMAK assessment; (b) INV score as previous, but reference n = 7 (inline image, dashed line) streams matched for INV score; (d) QMCI scores; (e) MCI scores.

Download figure to PowerPoint

HABITAT scores from treatment streams improved over the study period, but a less marked positive change was also seen in reference streams, so there was no treatment effect (F2,56 = 1·034, P = 0·362; Fig. 4c). Inclusion of the significant covariates DISCHARGE and SDSP (F2,54 = 9·20, P < 0·001 together) did not reveal evidence for a difference between changes in treatment and reference groups (F2,54 = 0·29, P = 0·747).

Changes in mean QMCI values in both groups over time (Fig. 4d) reflected those seen in INV scores (Fig. 4a), with an overall increase from March 2001 to 2003 in the treatment group, while the mean reference stream QMCI score gradually declined (F2,56 = 2·615, P = 0·082). This group effect was weaker than that seen for INV scores. SDSP was again a significant covariate in an ancova explaining variation in QMCI scores over time (F1,55 = 4·65, P = 0·036; for effect of all other covariates, minimum P = 0·616); when this was taken into account the group × time interaction effect on QMCI scores was somewhat reduced, but remained present (F2,55 = 2·48, P = 0·093).

MCI scores initially declined in both BMP-adopted and reference stream groups (Fig. 4e). While scores from BMP-adopted streams did improve between March 2002 and 2003, there was no significant between-group difference in the overall changes recorded (F2,56 = 0·662, P = 0·520). Inclusion of SDSP (stock days in study paddock, F1,55 = 4·60, P = 0·036; minimum P among all other background variables entered in ancova = 0·085) as a predictor did not reveal any effect of BMP adoption on the MCI (F2,55 = 0·34, P = 0·710).

Inclusion of the three streams in which management had not been changed until after March 2002 (Table 3) can be expected to have made these tests for effects more conservative. Note also that other individual parameters that appeared to drive differences within the ordinations were tested unvariately, but no trends were significant (P > 0·5 in all cases).

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

information effects on stream health

Investigation of environmental outcomes of information provision is a novel approach to the assessment of environmental education. The in-stream study presented here, together with a study examining the association between information provision and adoption of BMPs in overall farm management (Rhodes 2004), provides insights into the value of environmental information at both environmental and behavioural levels. Comparisons of stream health between ‘information’ and ‘no information’ streams were not significant, although there were encouraging patterns in overall stream health indices (SHMAK invertebrate and habitat scores) that are consistent with a beneficial effect of information. Despite our best efforts, our study design suffered from several constraints, including lack of prior knowledge of streams at which management would be changed (some farmers adopted BMPs, but on streams other than the one chosen for study) and the relatively short time-frame. Whether further integrated research of this type is warranted is debatable: while theoretically pleasing, it may be more useful to pursue separate studies of information–adoption and adoption–environment relationships. Such studies may identify more readily components of information or adoption that are important for achieving change, and still allow valid conceptual linking of results to predict information–environment outcomes. Thus, the behavioural study of Rhodes (2004) that showed farmers who received information were more likely to incorporate BMP into aspects of overall farm management within 18 months has now been associated with improvements to agricultural stream health within a period as short as 2 years (this study).

bmp adoption effects on stream health

Improvements in condition over 2 years were consistent among the seven BMP-adopted streams where management had been modified to improve stream health. Although no overall pattern was seen in streams where management was not changed, some of these reference streams showed changes similar to BMP-adopted streams, sometimes with equal or greater magnitude (e.g. streams 1, 11, 26 and 28; Fig. 3a,b). The differences observed were therefore not unique to streams where BMPs were adopted; that is, the changes recorded after 2 years are within the realm of natural variation or effects of routine management in agricultural streams. What is important is that these changes occurred consistently among BMP-adopted streams but not among reference streams.

The individual variable most responsible for differences among the two stream groups was INV. Like the more rigorous QMCI (Stark 1985, 1998), this biotic index is weighted for pollution sensitivity of taxa and relative abundances. Our finding of similar responses in both the QMCI scores and INV scores supports the validity of the lower-resolution index. Of the parameters assessed, an invertebrate community index is perhaps most likely to demonstrate a response to management change as it integrates multiple environmental factors over a longer time-frame than just the sampling occasion (de Pauw & Vanhooren 1983; Boothroyd & Stark 2000). While MCI scores have been related positively to protected stream areas (Storey & Cowley 1997; Scarsbrook & Halliday 1999) and have been correlated more highly with water quality than QMCI scores (Quinn & Hickey 1990), we found no evidence for BMP effects on MCI. It appears that the effect of the adopted BMPs on invertebrate communities was subtle, being evident only when using indices that take both invertebrate abundance and pollution-sensitivity into account.

The QMCI index was designed originally to detect organic enrichment. However, it can reflect other pollution types, including substrate fines (Boothroyd & Stark 2000). Although none of the six physico-chemical measures investigated (NOx-N, NH4-N, DRP, TSS, BARE and FINES) showed significant improvements in BMP-adopted compared to reference streams (results not shown), it is likely that a 6-monthly sampling regime is inadequate to quantify significant changes in these measures. Water quality variables may fluctuate on a daily basis (Vincent & Downes 1980; Sweeney 1993) and the attempts in this study to control for factors that might cause this, using the ‘background management’ covariates, are unlikely to have completely accounted for such variation. Our concurrent measurement of upstream water quality at the source of each stream, or its entrance to the property, gives us confidence that the lack of effect is not due to contamination from farms upstream. Previous studies have shown significant improvements in nutrients and suspended solids after riparian retirement using monitoring on a half-hourly or weekly basis (Howard-Williams & Pickmere 1994; Williamson et al. 1996).

factors that make this study conservative

Several limitations in study design make this a somewhat conservative evaluation of the effects of BMP-adoption. A time-frame of at least 3–5 years has been cited as necessary for documentation of water quality improvements associated with management change (Vought et al. 1994; Larsen et al. 1998), whereas our analyses cover only a 2-year period. Furthermore, inconsistency when different farmers adopted BMPs voluntarily (some BMPs were adopted only months before final evaluation) may have resulted in underestimation of the true magnitude of some effects. Variation in environmental conditions and daily or yearly farm management added to the complexity of the study. Problems such as varying stocking or fertilizer rates have been acknowledged in earlier studies (e.g. Williamson et al. 1996; Leeds-Harrison et al. 1999), but are rarely accounted for. Here, attempts were made to statistically account for the effects of some variables, but others (such as high discharge events) could not be controlled.

Given these limitations, it is notable that the positive effects shown were associated with relatively minor changes in management. Patterns in the present study suggest that, at relatively high stocking rates, simply moving stock from stream paddocks before visible streambank damage occurs may contribute to increases in INV and QMCI (streams 3 and 13). Even when streams were fenced, improvements were seen despite less-than-ideal protection – for example, continued lower intensity grazing at stream 8, retention of a short unfenced section for stock access to water at stream 17, and despite the fact that the sampling site at stream 21 was still grazed, being 235 m below the 150 m fenced section. It is not possible to assess whether the relatively small individual changes that contribute to the ‘treatment’ effect in this short-term study can achieve worthwhile outcomes over the long term, but significant improvements at this early stage are encouraging. This is relevant as these changes were chosen voluntarily by farmers and, presumably, represent the type of change that can be anticipated under New Zealand's current, largely voluntary agricultural stream protection system.

Conclusion

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

This ‘BACI’-type study evaluated short-term effects of provision of information and voluntary adoption of a range of fairly minor BMPs for stream health. No significant information–stream health relationships were detected. However, the positive relationship between receipt of information and adoption of BMPs into wider farm management (Rhodes 2004), and the relationship between BMP adoption and stream health on the same farms (this study), make it seem likely that a positive information–stream health association will be manifest with time. Despite difficulties, such as the inability to properly control reference stream conditions, and even though we used relatively simple methods to assess stream health that were designed for application by farmers, discernible improvements were demonstrated in streams where BMPs were adopted. Limited replication means the value of individual BMPs cannot be assessed, but positive outcomes in stream health soon after BMP implementation are encouraging. However, the magnitude of changes within individual treatment streams was no greater than that seen at some reference streams as a result of natural or rotation-grazing-induced variation. Long-term appraisals of BMP adoption, focusing on practices that are less management- and resource-intensive for farmers, are required to assess whether early changes, such as those documented here, will develop into environmentally useful improvements.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References

We thank the many farmers who gave access to their waterways and gave their time completing surveys, and also numerous field and laboratory assistants. This work was funded by a Ministry of Agriculture and Forestry Sustainable Farming Fund grant.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References
  • Australia and New Zealand Environment and Conservation Council (ANZECC) (2000) Australian and New Zealand Guidelines for Fresh and Marine Water Quality. Australia and New Zealand Environment and Conservation Council, Canberra.
  • American Public Health Association (APHA) (1998) Standard Methods for the Examination of Water and Wastewater, 20th edn. American Public Health Association, Washington DC.
  • Biggs, B., Kilroy, C. & Mulcock, C. (1998) New Zealand Stream Health Monitoring and Assessment Kit: Stream Monitoring Manual, version 1. Technical Report no. 40. National Institute of Water and Atmospheric Research, Christchurch.
  • Boothroyd, I. & Stark, J. (2000) Use of invertebrates in monitoring. New Zealand Stream Invertebrates: Ecology and Implications for Management (eds K.J.Collier & M.J.Winterbourn), pp. 344373. New Zealand Limnological Society, Hamilton.
  • Ter Braak, C.J.F. & Smilauer, P. (1998) canoco Reference Manual and User's Guide to canoco for Windows: Software for Canonical Community Ordination, version 4. Microcomputer Power, Ithaca.
  • Brougham, R.W. (1955) A study in rate of pasture growth. Australian Journal of Agricultural Research, 6, 804812.
  • Chapman, M.A. & Lewis, M.H. (1976) An Introduction to the Freshwater Crustacea of New Zealand. Collins, Auckland.
  • Collier, K.J., Fowles, C.R. & Hogg, I.D. (2000) Management, education and conservation. New Zealand Stream Invertebrates: Ecology and Implications for Management (eds K.J.Collier & M.J.Winterbourn), pp. 374400. New Zealand Limnological Society, Hamilton.
  • Cook, M.G., Hunt, P.G., Stone, K.C. & Canterberry, J.H. (1996) Reducing diffuse pollution through implementation of agricultural best management practices: a case study. Water Science and Technology, 33, 191196.
  • Cornforth, I.S. & Sinclair, A.G. (1984) Fertiliser and Lime Recommendations for Pastures and Crops in New Zealand. Ministry of Agriculture and Fisheries, Wellington.
  • Gale, J.A., Line, D.E., Osmond, D.L., Coffey, S.W., Spooner, J., Arnold, S.A., Hoban, T.J. & Wimberley, R.C. (1993) Evaluation of the Experimental Rural Clean Water Programme. National Water Quality Evaluation Project, NCSU Water Quality Group, Raleigh, NC and Biological and Agricultural Engineering Department, North Carolina State University, North Carolina.
  • Green, R.H. (1979) Sampling Design and Statistical Methods for Environmental Biologists. Wiley, Chichester.
  • Hawkes, H.A. (1998) Origin and development of the Biological Working Party score system. Water Research, 32, 964968.
  • Howard-Williams, C. & Pickmere, S. (1994) Long-term vegetation and water quality changes associated with the restoration of a pasture stream. The Restoration of Aquatic Habitats (ed. K.Collier), pp. 93109. Department of Conservation, Wellington.
  • Larned, S.T., Scarsbrook, M.R., Snelder, T.H., Norton, N.J. & Biggs, B.J.F. (2004) Water quality in low-elevation streams and rivers of New Zealand: recent state and trends in contrasting land-cover classes. New Zealand Journal of Marine and Freshwater Research, 38, 347366.
  • Larsen, R.E., Krueger, W.C., George, M.R., Barrington, M.R., Buckhouse, J.C. & Johnson, D.E. (1998) Livestock influences on riparian zones and fish habitats: literature classification. Journal of Range Management, 51, 661664.
  • Leeds-Harrison, P.B., Quinton, J.N., Walker, M.J., Sanders, C.L. & Harrod, T. (1999) Grassed buffer strips for the control of nitrate leaching to surface waters in headwater catchments. Ecological Engineering, 12, 299313.
  • De Pauw, N. & Vanhooren, G. (1983) Method for biological quality assessment of watercourses in Belgium. Hydrobiologia, 100, 153168.
  • Perry, C.D., Vellidis, G., Lowrance, R. & Thomas, D.L. (1999) Watershed-scale water quality impacts of riparian forest management. Journal of Water Resources Planning and Management, 125, 117125.
  • Quinn, J.M., Cooper, A.B., Davies-Colley, R.J., Rutherford, J.C. & Williamson, R.B. (1997) Land-use effects on habitat, water quality, periphyton and benthic invertebrates in Waikato, New Zealand, hill country streams. New Zealand Journal of Marine and Freshwater Research, 31, 579597.
  • Quinn, J. & Hickey, C.W. (1990) Characterisation and classification of benthic invertebrate communities in 88 New Zealand rivers in relation to environmental factors. New Zealand Journal of Marine and Freshwater Research, 24, 387409.
  • Quinn, J.M., Williamson, R.B., Smith, R.K. & Vickers, M.L. (1992) Effects of riparian grazing and channelisation on streams in Southland, New Zealand. II. Benthic invertebrates. New Zealand Journal of Marine and Freshwater Research, 26, 259273.
  • Rhoads, B.L., Wilson, D., Urban, M. & Herricks, E.E. (1999) Interaction between scientists and nonscientists in community-based watershed management: emergence of the concept of stream naturalization. Environmental Management, 24, 297308.
  • Rhodes, H.M. (2004) Downstream effects of stream management information on farms: an assessment of the behavioural and environmental outcomes of providing farmers with information about stream management. PhD Thesis, University of Otago.
  • Scarsbrook, M.R. & Halliday, J. (1999) Transition for pasture to native forest land-use along stream continua: effects of stream ecosystems and implications for restoration. New Zealand Journal of Marine and Freshwater Research, 33, 293310.
  • Sovell, L.A., Vondracek, B., Frost, J.A. & Mumford, K.G. (2000) Impacts of rotational grazing and riparian buffers of physicochemical and biological characteristics of Southeastern Minnesota, USA, streams. Environmental Management, 26, 629641.
  • SPSS Inc. (2001) SPSS for Windows, release 11·0·0. SPSS Incorporated, Chicago, IL.
  • Stark, J.D. (1985) A Macroinvertebrate Community Index for Stony Streams. Water and Soil Miscellaneous Publication 87. National Soil and Water Conservation Authority, Wellington.
  • Stark, J.D. (1998) SQMCI: a biotic index for freshwater macroinvertebrate coded abundance data. New Zealand Journal of Marine and Freshwater Research, 32, 5565.
  • Stark, J.D., Boothroyd, I.K.G., Harding, J.S., Maxted, J.R. & Scarsbrook, M.R. (2001) Protocols for Sampling Macroinvertebrates in Wadeable Streams. New Zealand Macroinvertebrate Working Group Report no. 1. Sustainable Management Fund Project no. 5103. Ministry for the Environment, Wellington.
  • Storey, R.G. & Cowley, D.R. (1997) Recovery of three New Zealand rural streams as they pass through native forest remnants. Hydrobiologia, 353, 6376.
  • Sweeney, B.W. (1993) Effects of streamside vegetation on macroinvertebrate communities of White Clay Creek in Eastern North America. Proceedings of the Academy of Natural Sciences of Philadelphia, 144, 291340.
  • Vincent, W.F. & Downes, M.T. (1980) Variation in nutrient removal from a stream by watercress (Nasturtium officinale). Aquatic Botany, 9, 221235.
  • Vought, L.B.M., Dahl, J., Pedersen, C.L. & Lacoursiere, J.O. (1994) Nutrient retention and processing in riparian ecotones. Ambio, 23, 342348.
  • Wang, S.W., Lassoie, J.P. & Curtis, P.D. (2006) Farmer attitudes towards conservation in Jigme Singye Wangchuck National Park, Bhutan. Environmental Conservation, 33, 148156.
  • Weigel, B.M., Lyons, J., Paine, L.K., Dodson, S.I. & Undersander, D.J. (2000) Using stream macroinvertebrates to compare riparian land use practices on cattle farms in Southwestern Wisconsin. Journal of Freshwater Ecology, 15, 93106.
  • Williamson, R.B., Smith, C.M. & Cooper, A.B. (1996) Watershed riparian management and its benefits to a eutrophic lake. Journal of Water Resources Planning and Management, 1, 2432.
  • Williamson, R.B., Smith, R.K. & Quinn, J.M. (1992) Effects of riparian grazing and channelisation on streams in Southland, New Zealand. 1. Channel form and stability. New Zealand Journal of Marine and Freshwater Research, 26, 241258.
  • Winterbourn, M.J., Gregson, K.L.D. & Dolphin, C.H. (2000) Guide to the aquatic insects of New Zealand. Bulletin of the Entomological Society of New Zealand, 13, 1102.
  • Withers, P.J.A. & Jarvis, S.C. (1998) Mitigation options for diffuse phosphorus loss to water. Soil Use and Management, 14, 186192.
  • Wolman, M.G. (1954) A method of sampling coarse river-bed material. Transactions of the American Geophysical Union, 35, 951956.
  • Zar, J.H. (1999) Biostatistical Analysis, 4th edn. Prentice Hall, New Jersey.