This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/authorsrights
Transcript
This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies areencouraged to visit:
Sharon R. Aarons a, *, Alice R. Melland b, Lianne Dorling a
a Future Farming Systems Research Division, Department of Environment and Primary Industries, Ellinbank Centre, 1301 Hazeldean Road, Ellinbank,Victoria 3818, Australiab Teagasc, Johnstown Castle Environment Centre, Wexford, County Wexford, Ireland
a r t i c l e i n f o
Article history:Received 7 February 2012Received in revised form23 August 2013Accepted 30 August 2013Available online 1 October 2013
Dairy farmers are encouraged to restrict stock access by fencing riparian zones to reduce streampollution and improve biodiversity. Many farmers are reluctant to create fenced riparian zones becauseof the perceived loss of productive pasture. Anecdotal reports indicate that pasture production in fencedareas is especially valued during summer months when water stress is likely to limit pasture growth inother areas of the farm. We measured pasture production, botanical composition, soil moisture, andfertility in Riparian (within 20 m of the riverbank), Flat (greater than 20 but less than 50 m from theriverbank), and Hill (elevated) areas on three commercial dairy farms from October 2006 to November2007 in south eastern Australia.
Riparian and Flat areas produced significantly more pasture, with on average approximately 25% moredry matter per ha grown in these areas compared with Hill paddocks. Percentage ryegrass was 14% loweron Hill slopes compared with Riparian and Flat areas and was compensated for by only a 5% increase inother grass species. Significant seasonal effects were observed with the difference in pasture productionbetween Hill, and Riparian and Flat areas most pronounced in summer, due to soil moisture limitationson Hill paddocks.
To examine potential productivity impacts of this lost pasture, we used a questionnaire-based surveyto interview the farmers regarding their farm and riparian management activities. The additional pasturethat would have been available if the riverbanks were not fenced to their current widths ranged from 6.2to 27.2 t DM for the 2006/2007 year and would have been grown on 0.4e3.4% of their milking area. If thispasture was harvested instead of grazed, the farmers could have saved between $2000 and $8000 oftheir purchased fodder costs in that year. By fencing their riparian areas to 20 m for biodiversity benefits,between 2.2% and 9.8% of their milking area would be out of production amounting to about $16,000 inadditional purchased fodder costs. We discuss the additional fencing, production, and on-going man-agement costs associated with fencing riparian areas, the costs to the environment and the enterprise ofstock freely accessing waterways, as well as the policy implications.
� 2013 Elsevier Ltd. All rights reserved.
1. Introduction
Farmers are expected to minimise pollutant movement fromdairy production systems to the environment, particularly to wa-terways; with changed riparian management one of the actionsencouraged both by the dairy industry and natural resource man-agers (see, for example, Wilcock et al., 2007). Riparianmanagementimprovements on land that is actively managed for dairy
production focus primarily on fencing land adjacent to waterwaysto exclude grazing cattle and to create buffer zones. Riparian filterstrips (buffer zones) have the potential to slow movement of pol-lutants (i.e., nutrients, sediment, and pathogens) into waterways(Dabney et al., 2006; Sullivan et al., 2007). Line et al. (2000) re-ported reduced emissions of nitrogen, phosphorus, and sediment towaterways when livestock were excluded from riparian areas,although the form of phosphorus may change to more bioavailableand potentially more hazardous forms (McKergow et al., 2003).Revegetating fenced riparian zones by planting native species oflocal provenance is also recommended to increase biodiversity inthese areas (Lovell and Sullivan, 2006). Greater biodiversity (nativesmall mammals, birds, and vegetation) was observed in fenced
* Corresponding author. Future Farming Systems Research Division, Departmentof Primary Industries, Ellinbank Centre, 1301 Hazeldean Road, Ellinbank, Victoria3821, Australia. Tel.: þ61 3 5624 2222; fax: þ61 3 5624 2200.
riparian land which had not been grazed for a minimum of sevenyears, in contrast to unfenced grazed areas (DPI, 2006).
The benefits of establishing fenced revegetated riparian buffersaccrue to both farmers and the wider community to various extentsand include better stock management and improved water quality(Aarons, 2011; Aarons and Gourley, 2012). However, farmer moti-vation to undertake conservation activities is influenced by anumber of factors (Beedell and Rehman, 1999; Greiner et al., 2009;Lankester et al., 2009), and a better understanding of factorsinfluencing farmers’ decisions may contribute to greater estab-lishment of riparian buffers (Lovell and Sullivan, 2006).
Previous studies of farmers’ motivations to undertake riparianmanagement reveal that production and financial factors areimportant in influencing adoption. Financial considerationsreduced the implementation of best management practices (suchas using fencing to manage stock access to creeks) for farmers innorthern Victoria (Curtis and Robertson, 2003). In northernAustralia riparian management practices that were perceived tohave a production benefit were most likely to be adopted(Lankester et al., 2009). Likewise, the reasons given by New Zealanddairy farmers for accepting or rejecting improved riparian man-agement practices were aligned to their primary farming goal ofensuring a viable farm business (Parminter and Pedersen, 2000).These authors report decreased stock losses and increased overallfarm viability as among the most frequent and most importantreasons (respectively) given for accepting a riparian managementpractice. Increased farm costs and weed problems as well as landwastage were primary reasons for farmer rejection of improvedriparian management practice (Parminter and Pedersen, 2000).Despite the positive relationship between access to informationand adoption of riparian management best practices, Rhodes et al.(2002) reported that “loss of productive land” was one of manyeconomic reasons cited by farmers as barriers to adoption ofimproved management of riparian zones. Using a decision treeapproach, Lynch and Brown (2000) report that land value and cropprice govern whether or not farmers are prepared to implementriparian buffers. An understanding of the costs associated withimproving riparian management could therefore contribute toincreased adoption by farmers of riparian fencing. Additionally,recommendations for public contributions to farmers need to bebased on accurate measures of the costs associated with fencingriparian areas (Lovell and Sullivan, 2006).
Loss of productive land, as perceived by farmers when riparianareas are fenced and revegetated, has not been quantified in theliterature. Farmers’ estimation of this ‘lost’ or ‘wasted’ land appearsto be based on calculations of the riparian area that would be fencedaswell as a perception of its relative pasture productivity in summercompared with other more elevated and drier parts of the farm.
To test the hypothesis that pasture produced in the riparianpaddocks is greater than that from the elevated paddocks of threegrazed dairy farms, pasture production and botanical compositionweremeasured on three commercial farms in the rain-fed Victoriandairy region of south eastern Australia. The potential impact onfarm productivity of pasture production excluded by fencing wasassessed for each dairy farm.
2. Materials and methods
2.1. Farm descriptions
Three commercial dairy farms, designated A, B, and C, wereselected based on the availability of suitable riparian sites and thewillingness of farmers to provide the required farm data and in-formation. The farms were located in the rain fed and predomi-nantly dairy region of West Gippsland in Victoria, Australia, where
Farms A and C were adjacent to each other while Farm B wasapproximately 30 km away. These farms, milking predominantlyFriesian dairy cows, were typical of local grazed dairy systems.Supplements such as silage, hay and grain were used to fill feedgaps in pasture availability throughout the year, where the pastureswere dominated by ryegrass (Lolium perenne L.) with white clover(Trifolium repens) forming less than 2% of pasture dry weight.Perennial species such as cocksfoot (Dactylis glomerata L.) andpaspalum (Paspalum dilatatum) and annuals (or short-lived pe-rennials) such as winter grass (Poa annua) were also present.Ryegrass constituted a mean of 71%, other perennial and annualgrasses, 21% and broad-leaf weeds 4% of pasture dry matter at thebeginning of this study.
2.2. Climate data
A temperate climate prevails in the region where spring typi-cally commences in September with summer following inDecember toMarch. Rains heralding the beginning of autumn occurfrom late March to early April and winter starts in June; with mostrainfall usually occurring in late winter/early spring. Based on re-cords collected for 120 years, Farm B received 133 mm less annualrainfall than Farms A and C (Table 1). However, more recentmeteorological data (1986e2008) gave a difference of 350 mmbetween the locations (Farms A and C, 1094 mm; Farm B, 744 mm).For the duration of this experiment however, drought conditionscontinued in the region with below average rainfall and greatertemperatures recorded (Fig. 1). Rainfall data were obtained eitherfrom farmers’ records for the year of the study or from datacompiled by one of the Australian Bureau of Meteorology weatherstations located adjacent to Farm B (BOM, 2012). Long term (1889e2008) continuous monthly climate data were obtained from‘patched point datasets’ (QCCCE, 2012), where interpolations areused to fill gaps in weather station data (see Jeffrey et al., 2001 asdescribed in QCCCE, 2012).
2.3. Sample design
Pasture production was measured in three topographical areasin selected paddocks on each farm: a riparian area (Riparian; within20 m of the riverbank), in riparian flats but outside the previouslydescribed riparian area (Flat; greater than 20 m but less than 50 mfrom the riverbank), and on elevated land as near to the riparianarea as was feasible and practical (Hill). The chosen paddocks weregrazed and managed throughout the experiment according to thefarmers’ normal practices. Pasture sampling commenced in eachtopographical area in the selected paddocks in spring 2006 andfinal measurements were made in November 2007. Prior to thecommencement of pasture sampling the topographical areas weresoil sampled for nutrient analysis. Each soil sample consisted of acomposite of a minimum of 30 cores of 2.5 cm diameter and 10 cmdepth.
Table 1Location and long-term average climate characteristics of the three dairy farms, A, B,and C.
Mean annual
Farms Latitude andlongitude
Elevation(m)
Tmaxa
(�C)Tmin
b
(�C)Rainfallc
(mm)Hill slope(%)
Hillaspect
A & C 38.25�S, 145.93�E 167 18.5 8.6 1050 2e5 NB 38.08�S, 146.21�E 161 19.2 8.8 917 10e15 E
a Average annual daily maximum temperature.b Average annual daily minimum temperature.c Average total annual rainfall.
S.R. Aarons et al. / Journal of Environmental Management 130 (2013) 255e266256
Five wire exclusion cages (each 1.4 � 2.1 m) were placedrandomly within each topographical area on each farm, withpasture and soil samples collected from within the cages and thecages representing replicate plots. The pasture in each plot wasmown to grazing height (5 cm) before the exclusion cages werepegged into position to prevent cows grazing the paddocksaccessing the pasture in the plots. Pasture beneath the cages on agiven farm was harvested at the 3-leaf stage (Fulkerson andDonaghy, 2001) to give an indication of the potential productivityof the pasture in the Riparian, Flat or Hill areas. After each harvest, anew set of five plots was established in each area, each at least 1.5maway from the location of the previous plots/cages. The exclusioncages were moved a minimum of eight times in each topographicalarea over the duration of the experiment (Spring 2006, Summer2006/2007, Autumn 2007,Winter 2007, Spring 2007, amaximum of14 months). No plot locations were re-sampled.
2.5. Sample collection and analysis
At each harvest pasture samples were collected for biomass andbotanical composition measurements and soils were cored tomeasure soil moisture. First a buffer zone was mown around eachplot before the exclusion cagewas removed. A quadrat (32� 64 cm)was then randomly thrown into the plot and all the pasture in thequadrat cut to 5 cm with hand-shears and collected for assessingbotanical composition. The remaining pasture was harvested to5 cm using a mower with a catcher. The combined weight of themown pasture and the pasture hand-harvested for botanical
composition was recorded. A sub-sample of the well-mixed mownpasture was collected for calculation of dry matter (DM) content.Three soil cores (2.5 cm diameter � 10 cm deep) were collectedfrom random locations in each plot for soil moisture calculations.
Pasture sub-samples were dried at 105 �C for 24 h to calculatethe pasture DM produced for each harvest. Daily pasture growthrates were calculated by dividing the pasture DM grown at eachharvest by the number of days in each harvest interval. The sameharvest interval was used for all topographical areas on each farm,but intervals differed fromharvest to harvest and from farm to farmbased on the farmer’s paddock management.
Of the five botanical composition samples collected, three wererandomly selected for sorting. Each sample was thoroughly mixed,quartered and the selected quarter further quartered until a mini-mum weight of 100 g was obtained. If the harvested botanicalcomposition samples each weighed less than 95 g fresh weight, forexample in summer, the three selected samples were not sub-sampled. Botanical composition samples were sorted intoryegrass and other perennial grasses, clovers, annual grasses,broadleaf weeds and dead plant material. The sorted plant materialwas dried at 105 �C for 24 h.
Soils collected for moisture content were dried at 105 �C for24 h. Soil samples for nutrient analysis were dried at 40 �C for 72 h,ground, then sieved to less than 2 mm. Soils were analysed for pHin water and 0.2 M CaCl2, electrical conductivity, total soluble salts,extractable aluminium, bicarbonate extractable phosphorus (Olsenet al., 1954) and potassium (Colwell, 1963), and calcium phosphateand charcoal extractable sulphur (Peverill et al., 1974). The phos-phorus buffering index of the soils in each topographical area wasalso calculated, based on the Olsen bicarbonate extraction. Soil
10
30
50
70
90
110
130
150
Jan Feb March April May June July Aug Sep Oct Nov Dec
Jan Feb March April May June July Aug Sep Oct Nov Dec Jan Feb March April May June July Aug Sep Oct Nov Dec
Rain
fall (m
m)
Farm A and C 2006 Farm A and C 2007 A&C Long term rain
a
10
30
50
70
90
110
130
150
Jan Feb March April May June July Aug Sep Oct Nov Dec
Farm B 2006 Farm B 2007 B Long term rain
b
0
5
10
15
20
25
30
Tem
peratu
re (o
C)
Farm A and C 2006 Farm A and C 2007 A&C Long term ave. temp
c
0
5
10
15
20
25
30
Farm B 2006 Farm B 2007 B Long term ave. temp
d
Fig. 1. Total monthly rainfall (mm) and average monthly temperature (�C) recorded for Farms A (a, c), B (b, d) and C (a, c) over the duration of the study compared with long termaverages for both locations. Average temperatures for the months of the experiment were interpolated from patch point datasets for weather stations adjacent to the study lo-cations. The long term average temperatures are the mean of the patch point data from 1891 to 2008.
S.R. Aarons et al. / Journal of Environmental Management 130 (2013) 255e266 257
Author's personal copy
samples were analysed according to the methods given in Raymentand Higginson (1992).
2.6. Farm data
To assess the impact of fencing on farm profitability a semi-structured questionnaire was developed and the farmers inter-viewed one-on-one. Farm data pertaining to milk production, sizeof the lactating herd, farm inputs such as off-farm feed supple-ments and nutrient use on the topographical areas and rest of thefarm were collected. In addition, information about the length ofcreek frontage, economic and other value of the land, as well asattitudes of these farmers to fencing riparian areas was collected inthese interviews (see Aarons, 2011).
2.7. Statistical analysis
All pasture and soil data were analysed to identify differencesassociated with farm, topographical area (i.e., Riparian, Flat, Hill)and season and to investigate relationships between pasturegrowth and soil moisture using Genstat 10.1 (Lawes AgriculturalTrust, VSN International Limited� 2007) for statistical analysis andSplus 8.0 (Insightful Corporation 2006) for graphical representa-tion. Pasture and soil water data were either log (pasture DMcontent, percent total other grasses) or square root (pasture DMproduction, pasture growth rate, percent dead plant material, soilwater) transformed for statistical analysis when a distinct patternof increasing variance with means was observed in the residualplots. The data were analysed using residual maximum likelihood(REML) analysis methods in Genstat. A mixed model was used withfixed factorial effects for farm by area-type by season (GenStat code,Farm*Season*Area), and random effects for harvests and areaswithin farms, and plots within areas (GenStat code, Farm.Harvest/Area/Plot). The unusually early start to summer in October 2006resulted in an unbalanced dataset as the spring 2006 data wascollected for only two of the three farms. Consequently these datawere discarded for this REML analysis. To further compare total DMproduction, production data were summed over harvests withinseasons for each of the 9 areas (one of each of 3 area types on eachof 3 farms), prior to analysis, as there were unequal numbers ofharvests per season, and unequal numbers of harvests per farm.The statistical mixed effects model for this analysis of DM pro-duction had fixed effects for farm by season by area-type (GenStatcode, Farm*Season*Area), and random effects for plots within areaand farm, split for season (GenStat code, Farm.Area.Plot/Season).Least significant intervals (LSI) at the P ¼ 0.05 significance levelwere calculated and used to plot error bars.
3. Results and discussion
3.1. Pasture production
Farm means for the pasture variables were statistically equiva-lent and not different; indicating that growth conditions on thesecommercial dairy farms were similar, despite the distance betweenlocations and differences in recorded annual rainfall. Significantlydifferent pasture DM production (P < 0.001), pasture growth rates(P < 0.001), and pasture DM content (P < 0.001), were recorded forthe topographical areas. Mean pasture DM produced per harvest onthe Riparian, Flat, and Hill areas were 1,846, 1,809, and 1463 kg DM/ha respectively. This equated to annual production of 14.0, 13.7, and11.1 t DM/ha year based on an average growth period of 48 daysbefore harvest. When DM productionwas analysed to sum the dataover harvests within seasons to account for the uneven numbers ofharvests per season, and unequal numbers of harvests per farm,
annual production on the Riparian, Flat and Hill areas were 15.3,14.1 and 11.6 t DM/ha year respectively, similar to that obtained inthe initial analysis. Farm pasture production in this region histori-cally ranged from 6.5 to 9.7 t DM/ha year, under lower fertilityconditions although up to 12.6 t DM/ha year was recorded underexperimental conditions (Doyle et al., 2000).
Despite the comparable farm means, significant (P < 0.001)farm � area interactions were observed for pasture DM productionand growth rate, but not for pasture DM content (P ¼ 0.102). Thesignificant interactions are reflected in the greater pasture DMproduced in Riparian and Flat areas compared with Hill paddockson Farms A and B, in contrast to the similar pasture production forall areas for Farm C (Fig. 2a) However, the only significant differencein pasture DM content was observed on Farm B where the Hillpasture DM was greater than that from the Riparian or Flat areas(Fig. 2b).
Pasture growth rates are influenced by how efficiently solarradiation is converted to carbohydrates and the effect of leaf areaindex on the fraction of solar radiation intercepted by the pasturecanopy. Consequently pasture production is influenced by, amongstother things, species composition, edaphic factors such as soilfertility and soil moisture, as well as seasonal factors as governed bythe climate (Pearson and Ison, 1997).
3.2. Botanical composition
Percentage ryegrass and percentage other grass species(P ¼ 0.032) were significantly different in the three topographicalareas. Ryegrass comprised an average of 67.9% of the DM producedin Riparian (66.7%) and Flat (69.1%) areas compared with 53.8% onHill areas (s.e.d ¼ 4.5). Increases in other grass species from anaverage of 8% on these areas (Riparian, 9.1%; Flat, 7.5%) to 13.7% onHill paddocks compensated in part for the lower percentageryegrass on Hill areas. On these farms the other grass speciesconsisted primarily of perennial species including paspalum,cocksfoot, couch (Cyndon dactylon), and annual grasses such as Poa,soft pigeon grass (Setaria viridis), and summer grass (Digitariasanguinalis). The predominance of annual grasses could havecontributed to the lower pasture production in Hill areas asobserved by Stockdale (1983).
While significant farm � area interactions were observed forpercentage of other grasses (P < 0.001) species, these interactionswere not significant for percentage ryegrass (P ¼ 0.056). Signifi-cantly more ryegrass was only observed on the Riparian and Flatareas of Farm A (Fig. 3a); with lower amounts of other grass speciesin these areas, although similar trends were observed on Farm C(Fig. 3b). Broadleaf weeds comprised a significantly larger amountof the pasture sward on Hill areas of Farm B unlike the other farmswhere the trend was for fewer broadleaf weeds in these areas(Fig. 3c): explaining the higher DM content of Farm B Hill pasture.Thus farmer perceptions of negative pasture production impactsassociated with fencing riparian areas are borne out by this data.
3.3. Soil moisture and fertility
Soil moisture content will strongly influence pasture DM pro-duction with, for example, pasture growth halted when soil volu-metric water content (q) falls below 0.20 (Moir 1994 as given in(Moir et al., 2000; Pearson and Ison, 1997). In this study soilmoisture was similar on all farms potentially explaining the similarpasture DMproduced on the farms. However, significant (P< 0.001)area and farm � area soil water content means were recorded. Soilwater contents were significantly lower on the Hill paddocks ofFarms B and C than the Flat and Riparian areas. In contrast, the soilwater content of the Hill soils on Farm A tended to be unexpectedly
S.R. Aarons et al. / Journal of Environmental Management 130 (2013) 255e266258
higher than that of the Riparian and Flat areas (Fig. 2c). Ripariansoils are expected to have greater water holding capacity due totheir proximity to the waterway and the lateral and verticalmovement of water in these zones (Malanson, 1993; Naiman andDécamps, 1997). The higher soil moisture would contribute togreater pasture growth during the drier months of the year(Marques da Silva et al., 2008). In addition, the Hill soil type onFarm A, classified as a Ferrosol (Isbell, 2002), has a higher hydraulicconductivity than the Riparian/Flat soil type, which is a Hydrosol.Despite the unexpectedly higher soil moisture of the Hill soils, lesspasture grew compared to the Riparian and Flat areas, possibly dueto the greater soil P fertility of the latter soils (Table 2).
No consistent trends in soil properties were observed betweenthe Hill and the Riparian and Flat pasture soils. In general the
latter soils were more similar than the Hill soils on all farms, asthese samples were collected from different parts of the samepaddock on each farm. Farm B Hill soils appeared to be lower inmost soil chemical properties except for soil P (Olsen and Colwell),with no difference in P buffering index. The opposite was true forFarm C, while Farm A Riparian and Flat soils appeared be higher inplant available P but lower in extractable soil K compared with theHill soils on that Farm. Although soil Al was greater in Riparianand Flat areas the levels were not considered toxic to pasturegrowth. The soil S values for most soils on the three farms werewell in excess of the recommended value of 8 mg/kg (Gourleyet al., 2007).
Nutrient inputs to the Farm A paddocks included annual appli-cations of fertiliser, excreta returned as the cows rotationally grazed
1000
1200
1400
1600
1800
2000
2200
2400
2600
A - R
ipar
ian
A - F
lat
A - H
ill
B - R
ipar
ian
B - F
lat
B - H
ill
C -
Rip
aria
n
C -
Flat
C -
Hill
Past
ure
dry
mat
ter p
rodu
ced
(kg
/ ha)
a
0.15
0.20
0.25
0.30
A - R
ipar
ian
A - F
lat
A - H
ill
B - R
ipar
ian
B - F
lat
B - H
ill
C -
Rip
aria
n
C -
Flat
C -
Hill
Past
ure
dry
mat
ter c
onte
nt (g
/ g) b
0.1
0.2
0.3
0.4
0.5
A - R
ipar
ian
A - F
lat
A - H
ill
B - R
ipar
ian
B - F
lat
B - H
ill
C -
Rip
aria
n
C -
Flat
C -
Hill
Soil
wat
er c
onte
nt (g
/ g
soil)
c
Fig. 2. Mean pasture dry matter production (a; kg DM/ha), mean pasture dry matter content (b; g DM/g fresh weight) and mean soil water content (c; g water/g soil) in Riparian,Flat, and Hill topographical areas for all harvests on the three dairy farms (A, B, C). Error bars are the least significant intervals for means at the P ¼ 0.05 significance level.
S.R. Aarons et al. / Journal of Environmental Management 130 (2013) 255e266 259
the pastures and nutrient additions from unconsumed hay andsilage fed out in the paddocks. Greater removal by the farmer offodder from the Hill (3.3 t DM/ha) compared with the Riparian/Flat(2.7 t DM/ha) paddock and larger returns in excreta to the Riparian/Flat paddocks used to hold cows overnight after the eveningmilking could have contributed to the greater soil P fertility of theRiparian and Flat paddock soil. Soil fertility and soil moisture bothaffect pasture production (Moir et al., 2000), such that yield isgreater when neither are limiting. The Hill Ferrosols in this studyare high in free iron oxide and strongly sorb P, as indicated by thehigh P buffering index value (Table 2), and have much lower ‘plantavailable’ P, which would restrict plant growth at low soil moisture,although species composition could also influence growth rate.Although soil N measurements weren’t made, the larger excretal
returns of N could also have contributed to increased pasturegrowth in the Riparian and Flat paddock.
Marques da Silva et al. (2008) reported bigger crop yields inlower parts of the landscape in average and drier years, which wereassociated with the relatively higher soil moisture in these areas.Similarly in this study, Riparian and Flat areas had greater pastureproduction and higher soil water contents; conditions which wouldhave been influenced by the drought prevailing in the region formuch of the duration of this experiment. Consequently theincreased pasture production in Riparian areas would have beenless obvious in non-drought years. Also, the tendency for theaccumulation of water in lower lying areas in wet years to restrictcrop growth (Marques da Silva et al., 2008) implies that pasturegrowth could be slowed in Riparian topographical areas in higher
40
4550
55
6065
70
75
8085
90
A - R
ipar
ian
A - F
lat
A - H
ill
B - R
ipar
ian
B - F
lat
B - H
ill
C -
Rip
aria
n
C -
Flat
C -
Hill
Perc
enta
ge ry
egra
ss (%
)
a
0
10
20
30
40
50
60
70
80
A - R
ipar
ian
A - F
lat
A - H
ill
B - R
ipar
ian
B - F
lat
B - H
ill
C -
Rip
aria
n
C -
Flat
C -
Hill
Perc
enta
ge o
ther
gra
ss s
peci
es (%
)
b
-5
0
5
10
15
20
25
A - R
ipar
ian
A - F
lat
A - H
ill
B - R
ipar
ian
B - F
lat
B - H
ill
C -
Rip
aria
n
C -
Flat
C -
Hill
Perc
enta
ge b
road
leaf
wee
ds (%
) c
Fig. 3. Mean percentage ryegrass (a), mean percentage other grass species (b) and mean percentage broadleaf weeds (c) in Riparian, Flat, and Hill topographical areas for all harvestson the three dairy farms (A, B, C). Error bars are the least significant intervals at the P ¼ 0.05 significance level.
S.R. Aarons et al. / Journal of Environmental Management 130 (2013) 255e266260
rainfall years. The relative production benefit of Riparian areas istherefore only likely to occur in dry years.
Perennial ryegrass is less drought-tolerant than other pasturegrass species such as cocksfoot and phalaris (Cunningham et al.,1994), explaining the lower percentage ryegrass observed on Hillslopes compared with Flat and Riparian areas (Fig. 3). WhileMarques da Silva et al. (2008) reported a greater proportion ofgrasses on slopes compared with valleys, they did not distinguishbetween grass species.
3.4. Seasonal factors
Seasonal differences in pasture production have also been re-ported, with lower production observed in the winter months forperennial species (Cunningham et al., 1994; Hennessy et al., 2008;Stockdale, 1983); due in part to the relationship between thetemperature index of pasture species and mean daily temperature(Pearson and Ison, 1997). Herbage growth by temperate pasturespecies at the lower daily temperatures common inwinter could beten percent of that at optimum growing temperatures. Similarseasonal variations in pasture growth rates were reported by Mootet al. (2009) for New Zealand pastures. As expected significantseason effects were observed for pasture DM produced (P ¼ 0.002),pasture growth rate (P ¼ 0.003), pasture DM content (P < 0.001),and soil moisture (P ¼ 0.004), but surprisingly not for percentageryegrass, other grass species, broadleaf weeds, or dead material.Highly significant (P < 0.001) area � season interactions wereobserved for the pasture production variables above, as well as forpercentage ryegrass (P¼ 0.003), and percentage other grass species(P ¼ 0.008) and significant interactions for soil water content(P ¼ 0.037).
As a consequence of unusually high temperatures and the lowrainfall experienced over the study period, summer growth con-ditions commenced earlier (October 2006) and continued well intoApril 2007 (Fig. 1). Significantly less DM was produced on the Hillpaddocks than on the Flat and Riparian topographical areas insummer, with a trend to lower Hill production in autumn also(Fig. 4a). Pasture growth rates were similar although themagnitudeof the differences between winter and spring were much greater(Fig. 4b). No significant differences in pasture production betweenthe three topographical areas were observed in winter or spring.This observation agrees with anecdotal reports by farmers that ri-parian areas are especially valued for their additional pastureproductivity over the dry summer period, and with seasonal
variations in pasture growth rates observed in Australia and NewZealand (Doyle et al., 2000; Moot et al., 2009). Pasture DM contentswere highest for all three topographical areas in summer, followedby autumn. However, only the DM contents of pastures harvestedfrom the Hill paddocks in summer were significantly greater thanthat from the Riparian or Flat areas (Fig. 4c).
In contrast to pasture DM contents, soil water contents of theHill paddocks were almost always less than that of the Flat or Ri-parian areas, with only the autumn Hill, Flat, and Riparian areashaving similar soil moisture (Fig. 4d). Soil moisture in winter wasabout 2 ½ times greater than that in summer, translating into athree-fold increase in pasture DM produced on the Hill paddock inwinter compared with summer. Winter and summer pasture pro-duction on the Riparian and Flat areas were almost equivalenthowever, despite mean winter temperatures and solar radiationbeing approximately 8 �C and 12 MJ/m2 day (respectively) lowerthan that in summer. The lower mean winter temperatures equateto a temperature index of approximately 0.5, or half the potentialpasture DM production at the more optimal summer temperatures(Pearson and Ison, 1997). Pasture production in summer wastherefore likely to be limited by soil water, more so on the Hill thanRiparian and Flat pastures. The significantly greater pasture DMcontents from the Hill areas in summer are indicative of the lowersoil moisture contents in these areas. Increasing day length andtemperatures also contributed to pasture production, withconsiderably greater DM grown in spring compared with winter.
Pasture growth rate was generally positively related to soilmoisture contents for all farms in spring 2006 and summer 2006/2007 as expected (Fig. 5). Growth rates in autumn were very low,most likely due to soil moisture limitations; while in winter highsoil moisture and low temperatures would have restricted pasturegrowth. The slightly greater growth observed in the second winterpasture harvest on Farm A would have been due to increasing soiltemperatures with the approach of spring. Pastures harvested inspring 2007 occurred after cages were re-established in late winterwhen soil moisture was high and temperatures cool. Consequentlyon Farms B and C soil water contents were very high especially onthe Flat areas of these farms.
While soil moisture is likely to be a major determinant ofpasture growth, solar radiation, air temperature and evaporationalso explain seasonal pasture growth rates, where other edaphicfactors are equal. A linear regression relationship between growthrate (kg DM/ha day; square root transformed), soil water, solarradiation and mean air temperature accounted for 61% of the
Table 2Soil nutrient characteristics collected at the commencement of the experiment for the Riparian, Flat, and Hill topographical areas on each dairy farm (A, B, and C).
Farm e ‘topographical’area
ECa
dS/mTSSb
%pH(CaCl2)
pH(H2O)
Al (KCl)c
mg/kgPhosphorus PBIf
(Olsen)K (Colwell)g
mg/kgS (CPC)h
mg/kgPd (Olsen) mg/kg Pe (Colwell) mg/kg
A e Riparian 0.13 0.04 4.5 5.1 77 45 130 250 210 9A e Flat 0.11 0.04 4.5 5.1 83 46 130 230 170 8A e Hill 0.15 0.05 4.8 5.3 35 20 82 660 190 19B e Riparian 0.21 0.07 5.3 5.8 10 18 66 160 160 24B e Flat 0.31 0.11 5.1 5.5 10 18 66 190 200 35B e Hill 0.15 0.05 4.9 5.4 11 32 94 170 89 15C e Riparian 0.09 0.03 4.4 4.9 10 34 110 250 130 9C e Flat 0.13 0.04 4.4 4.9 10 28 92 220 190 15C e Hill 0.27 0.09 4.5 5 10 66 230 370 400 27
a Electrical conductivity.b Total soluble salts calculated as given in Rayment and Higginson (1982).c Plant available aluminium (Al) extracted in KCl.d Plant available phosphorus (P) extracted according to the Olsen et al. (1954) method.e Plant available P extracted according to the Colwell (1963) method.f P buffering index using the Olsen et al. (1954) extraction.g Plant available potassium (K) extracted according to the Colwell (1963) method.h Plant available S extracted by the calcium phosphate and charcoal (CPC) method (Peverill et al., 1974).
S.R. Aarons et al. / Journal of Environmental Management 130 (2013) 255e266 261
variability in the data (Equation (1)). However 49% of that vari-ability was attributed to solar radiation. Solar radiation and meanair temperatures were calculated using interpolated data from thepatch point datasets for each pasture growth interval.
In contrast, soil moisture accounted for more than 60% of sitevariability in pasture production in British grasslands, despite theimportance of temperature and light (Lazenby, 1988).
Percentage ryegrass was generally similar for all seasons excepton Hill paddocks in winter 2007 where ryegrass was significantlylower (Table 3). Increases in other grass species and broadleafweeds compensated for the reduced percentage ryegrass. SummerRiparian pastures tended to have less ryegrass and significantlymore grasses other than ryegrass which could contribute to loweroverall pasture quality and potential decreases in milk production(Stockdale, 1999). More dead plant material was recorded on Hillslopes during the summer period which would have contributed tothe greater percent DM recorded compared to the riparian and flatareas. Not surprisingly, a tendency for greater dead material insummer compared with other seasons (P ¼ 0.073) was observed.Thus, milk production could have been influenced in this droughtyear by the lower soil moistures and higher temperatures in sum-mer contributing to greater dead plant material on hill slopes incomparison to non-drought years.
3.5. Impact on farm productivity
Farmer motivation for, and subsequent adoption of, changedpractice determines the extent to which riparian management isimproved in dairy landscapes (Aarons, 2011). Impacts on
profitability can be an important driver for many farmers making itrelatively disadvantageous to implement new practices (Pannellet al., 2006). Thus production-specific information could meet thebusiness-related “aspirations” and “frames of reference” of farmers,thereby increasing their motivation to undertake improved riparianmanagement (Boxelaar and Paine, 2005; Lankester et al., 2009;Parminter and Nelson, 2003).
Having quantified the pasture produced in riparian and flatareas relative to elevated parts of the farm, we estimated the valueto each of the study farmers in the 2006/2007 lactation of notfencing their riverbank to their current width (Table 4). The fencedareas comprised from 0.4 to 3.4% of the farmers’ milking area andthe pasture that would have been available ranged from 6.2 to27.2 t DM/y. The farmers could have elected to milk more cows toutilise the additional pasturewhich, assuming no change in pastureutilisation, would have increased their herd by one to two cows. Incontrast, these farmers could have milked the same number ofcows and reduced the amount of brought-in feed purchased. In thislatter instance the farmers would have reduced their brought-infodder costs by $2000 to $8000 or between 1 and 6% of theircosts of purchasing fodder for the drought year of 2006/2007.
Various authors have suggested that farmer decisions regardingriparian management activities are often influenced by the avail-ability of financial incentives (Lynch and Brown, 2000; Rhodeset al., 2002), although Rhodes and colleagues suggest that the in-centives need be a realistic measure of the financial outlay byfarmers. The analysis undertaken in our case study does not includethe value of the three study farmers’ land or their fencing and la-bour costs. However, these costs were described by the farmers andestimated by one study farmer (Aarons, 2011). In Victoria, manyfarmers lease riparian land from the state government at a nominalcost and do not own the land that is to be fenced. In this instance, aconcern of farmers is the cost incurred to lease land that they nolonger use for production purposes, but often must manage. Otherexpenses that could be factored in are those associated with usingand managing riparian land for production purposes (Frimpong
0
1000
2000
3000
4000
Rip
aria
n Su
2006
/07
Flat
Su
2006
/07
Hill
Su20
06/0
7
Rip
aria
n Au
2007
Flat
Au
2007
Hill
Au 2
007
Rip
aria
n W
i20
07
Flat
Wi 2
007
Hill
Wi 2
007
Rip
aria
n Sp
2007
Flat
Sp
2007
Hill
Sp 2
007
Past
ure
dry
mat
ter (
kg /
ha) a
0
20
40
60
80
100
Rip
aria
n Su
2006
/07
Flat
Su
2006
/07
Hill
Su20
06/0
7
Rip
aria
n Au
2007
Flat
Au
2007
Hill
Au 2
007
Rip
aria
n W
i20
07
Flat
Wi 2
007
Hill
Wi 2
007
Rip
aria
n Sp
2007
Flat
Sp
2007
Hill
Sp 2
007
Past
ure
grow
th ra
te (k
g D
M /
ha. d
ay)
b
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
Rip
aria
n Su
2006
/07
Flat
Su
2006
/07
Hill
Su20
06/0
7
Rip
aria
n Au
2007
Flat
Au
2007
Hill
Au 2
007
Rip
aria
n W
i20
07
Flat
Wi 2
007
Hill
Wi 2
007
Rip
aria
n Sp
2007
Flat
Sp
2007
Hill
Sp 2
007
Past
ure
dry
mat
ter c
onte
nt (g
/ g) c
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
0.55
Rip
aria
n Su
2006
/07
Flat
Su
2006
/07
Hill
Su20
06/0
7
Rip
aria
n Au
2007
Flat
Au
2007
Hill
Au 2
007
Rip
aria
n W
i20
07
Flat
Wi 2
007
Hill
Wi 2
007
Rip
aria
n Sp
2007
Flat
Sp
2007
Hill
Sp 2
007
Soil
wat
er c
onte
nt (g
/ g
soil) d
Fig. 4. Mean pasture dry matter production (a; kg DM/ha), mean pasture growth rate (b; kg DM/ha.day), mean pasture dry matter content (c; g DM/g fresh weight), and mean soilwater content (d; g water/g soil) in the Riparian, Flat, and Hill topographical areas of all farms, over four of the seasons (summer, Su 2006/07; autumn, Au 2007; winter, Wi 2007;spring, Sp 2007) of this study. Error bars are the least significant intervals for means at the P ¼ 0.05 significance level.
S.R. Aarons et al. / Journal of Environmental Management 130 (2013) 255e266262
Author's personal copy
et al., 2007). These scientists suggest that riparian areas prone toflooding are more prevalent as stream order increases, are likely tohave lower land values and consequently are of lower value to theenterprise when excluded from farm production. Thus riparianland associated with lower stream orders, which can be the mostimportant for water pollution mitigation (Weaver et al., 2001),would have higher land values that would need to be considered incalculations of the cost to dairy enterprises of stock exclusion fromwaterways. On this basis, the estimated value of land wouldeffectively increase for Farm B with Farm A marginally greater thanFarm C.
As well as accurate estimates of farmer financial outlay, Nanereet al. (2007) recommend that calculations of production costs ofimproved natural resource management need to include those
associated with the environmental damage should improvedmanagement not be adopted. The environmental impact due tostock accessing waterways includes, primarily, nutrient and path-ogen inputs from excreta and sediment additions due to erosion ofriverbanks. These inputs degrade water quality on-farm as well asthat lower in the catchment. On-farm impacts of accessing poorquality water are likely to include increased animal health costs andreduced milk production (Castro-Hermida et al., 2009). Loss of landdue to continued erosion of riverbankswould be another direct costto the enterprise. These expenses need to be calculated andincluded in estimates such as thosemade for these farms. Only thencan more precise evaluation of the financial outlay required byfarmers to fence riparian zones be made.
Off-farmwater pollution contributes to algal bloomswhich are acost to recreation and fisheries industries, and negatively affectaquatic and terrestrial biota. These are largely reduced when stockare excluded from waterways by fencing, irrespective of the widthof the fenced zone. In fact, Dabney et al. (2006) suggest than anybuffer width is better than no buffer. However, Dorioz et al. (2006)recommend that buffer widths should be based on contaminantsource area contributions, topography and rainfall intensity, whichmay vary along the length of thewaterway. Despite this farmers areencouraged by natural resourcemanagers to fence to aminimum of10 m from the riverbank, although 20 m is preferred (MelbourneWater Stream Frontage Management Program, (Melbourne Water,2012), and required before farmers receive the highest cost shareof 80%. Likewise, strips of 30e50 m are recommended on the ma-jority of waterways on farms in Victoria to reduce nutrient andsediment losses into waterways (Straker and Lowe, 2004), whileproviding wildlife corridors and habitat patches of an adequate sizefor native species particularly small mammals and birds. Recom-mendations to farmers to improve biodiversity assets not onlyinclude fencing riparian zones, but also require revegetation ofthese areas with native vegetation of local provenance.
We calculated the production costs associated with setting aside20 m riparian zones on the three commercial dairy farms in this
Table 3Mean percentage ryegrass, mean percentage other grass species, mean percentagebroadleaf weeds, and mean percentage dead plant material harvested from plots inthe Riparian, Flat, and Hill topographical areas on three dairy farms during the fourseasons (Summer 2006/2007, Autumn 2007, Winter 2007, and Spring 2007) of theexperiment.
Season Area Composition of each plant component (%)
a Denotes a statistically significant least significant interval (P ¼ 0.05 level), foreach plant component between areas and seasons.
0.0 0.3 0.6 0.0 0.3 0.6
0.0 0.3 0.6 0.0 0.3 0.6 0.0 0.3 0.6Soil water content (g water / g soil)
60
140
60
140
60
140
Past
ure
grow
th ra
te (k
g D
M /
ha.d
y)Spring 2006 Summer 2006/7 Autumn 2007 Winter 2007 Spring 2007
Spring 2006 Summer 2006/7 Autumn 2007 Winter 2007 Spring 2007
Spring 2006 Summer 2006/7 Autumn 2007 Winter 2007 Spring 2007
A A A A A
B B B B B
C C C C C
RiparianFlatHill
Fig. 5. Pasture growth rate (kg DM/ha. day) and soil water content (g water/g soil) of Farms A, B and C at each harvest for each of the five plots in the Riparian, Flat and Hilltopographical areas during Spring 2006, Summer 2006/2007, Autumn 2007, Winter 2007 and Spring 2007. No samples were collected from Farm C in Spring 2006.
S.R. Aarons et al. / Journal of Environmental Management 130 (2013) 255e266 263
Author's personal copy
study using the assumptions given previously (Table 4). The per-centage of the milking area that would be taken out of productionranged from 2.2 to 9.8%, amounting to between 59 and 78 t pastureDM/yr farm (Table 4). The reduced availability of pasture for grazingcould be expected to have an impact on the farmers’ income due toa requirement to decrease herd sizes by 6e7 cows (that is, 1e5% ofthese herds). A net decline in farm income of between 1 and 3%wascalculated by Westra et al. (2005) when they used a modellingapproach to estimate the impact on farm income of implementa-tion of best management practices to reduce suspended sedimentin two catchments. The practices implemented in the study byWestra et al. (2005) included 30 m riparian buffer strips as well asnutrient and tillagemanagement changes. If on the other hand herdsizes were maintained in this study, additional brought-in foddercosts averaging $16,000 would be required to make up the pasturedeficit, resulting in an additional expenditure of between 4 and 12%of these farmers’ fodder costs for this lactation.
The expenses calculated do not account for the costs to thesefarmers of any additional water points that may be required, or thetime and costs associated with weed management and mainte-nance of areas fenced at widths for enhanced biodiversity benefit;considerations that would influence farmers’ perceptions of net
benefits associated with adopting this practice (Pannell et al.,2006). The issues associated with weed management areacknowledged by these and other farmers as a considerabledisincentive (Aarons, 2011; Curtis and Robertson, 2003). As withthe previous perceptions of lost pasture productivity, there iscurrently no credible data quantifying the on-going managementcosts (i.e., labour, replacement of plant material, weed manage-ment) of maintaining fenced riparian areas of different widths.Pannell et al. (2006) suggest that adoption of practices on a largescale will only occur infrequently if these practices are consideredunprofitable by farmers. In addition to environmental goals thefarmers in our study acknowledged production benefits associatedwith fencing riparian areas, such as improving herd and paddockmanagement (Aarons, 2011). Future research into on-going costs isrequired to provide needed information for farmers and naturalresource managers, and also to inform policy development, thuscontributing to increasing farmer adoption of improved practice(Greiner et al., 2009).
Where farmers are minimising the impact of their enterprise onthe environment, the costs for implementation of best practice, andthus their ‘duty of care’, can be deemed essentially their re-sponsibility (Cocklin et al., 2006). While these have yet to bequantified, providing any production-related benefits of adoptingriparian best practice, in addition to financial incentives and tar-geted education programs, could encourage practice change(Greiner et al., 2009; Lankester et al., 2009; Pannell et al., 2006;Rhodes et al., 2002). In contrast, in undertaking improved ripar-ian management practices targeted to increasing native biodiver-sity, the case could be made that the benefits largely accrue to thecommunity who should contribute to defraying expenses incurredby farmers (Curtis and De Lacy, 1996). Unless farmers can be shownto incur additional ecosystem services benefits from managingthese riparian areas for biodiversity conservation and theecosystem services values are quantified, economic policy in-struments whereby farmers are ‘paid’ for largely public benefitsmay need to be developed. However farmer ‘willingness to be paid’is influenced by many factors which need to be identified andquantified (Patrick et al., 2009). In so doing, increased adoptionmayoccur as appropriate policies are developed and community sup-port for farmer action is demonstrated (Greiner et al., 2009;Vanclay, 2004).
4. Conclusions
In this study pasture production was approximately 25% greaterin riparian (Riparian and Flat) areas than that on non-riparian (Hill)paddocks of these grazed dairy farms, and was dominated byperennial ryegrass. The lower pasture biomass grown on Hill pad-docks in summer compared with Riparian and Flat areas appearedto be due to generally lower soil moisture in the elevated areas. Incontrast the higher soil moisture of the Riparian and Flat areasappeared to limit growth rates in winter. Despite the greaterpasture production observed in riparian areas in this drought year,the relatively greater riparian pasture growth would most likely beless in average rainfall years and could even be further inhibited inwet years. Consequently access to riparian pasture is unlikely to bea benefit in all years.
Productivity on the three farms would have been marginallyaffected if the farmers had access to currently fenced riparian landbased on the riparian pasture production recorded in this study.The ability to milk an additional one or two cows or reducingpurchased fodder by between 0.5 and 6.4% is not considered asignificant addition to farm income and needs to be consideredwithin the context of the environmental and production costs ofnot excluding stock from waterways. Should these farmers fence
Table 4Farm productivity, riparian area, and impact on production data for the three dairyfarms.a
Farm A Farm B Farm C
Farm productivityFarm area (ha) 230 75 212Value of landb ($/ha) 12,000þ 20,000 w10,000Milking area (ha) 180 62.96 118Size of lactating herd 480 143 345Stocking rate (cow/ha) 2.7 2.3 2.9Milk production(L; [2006/2007])
2.75 � 106 800,000 2 � 106
Fodder costs ($) 420,000 127,289 352,950Riparian areaWaterway frontage (km) w2 3.1 2.3Proportion fenced All All 70%Width of fenced area (m) 5 5e10 for most <4Average pasture growth ratec
(kg DM/ha. dy)48.63 34.45 34.88
Impact on production when:Not fenced to current widthsFenced aread (ha) 1 2.16 0.48Percent of milking area 0.56% 3.43% 0.41%Pasture that would have been
Fenced to 20 mLand area (ha) 4 6.17 4.6Percent of milking area 2.22% 9.81% 3.90%Pasture lost (t DM/yr) 71 77.63 58.56Adjusted herd size 475 138 340Additional fodder costsf ($) 15,975 15,1389 15,724Costs (% expended) 3.8% 11.9% 4.5%
a Farm productivity and riparian management data (except pasture growth rate)provided by farmers during one-on-one interviews (Aarons, 2011).
b Land value estimated by the farmer. All dollar amounts are given as Australiandollars.
c Riparian average daily pasture growth rates from statistical analysis of datacollected in this study.
d Fenced area based on length of waterway frontage and width of current fencing.For Farms B and C, widths of 7 m and 3 m respectively were used.
e Herd sizes adjusted to account for additional fodder grown if riparian areas oneach farm were not fenced; based on 50% pasture utilisation and an estimated5.5 t DM required for each lactating cow.
f Cost of unnecessarily purchased fodder if riparian area is not fenced to currentwidths; or costs of additional brought in fodder required to compensate for lostpasture in 20 m riparian fenced areas. Calculations based on $300/t of hay.
S.R. Aarons et al. / Journal of Environmental Management 130 (2013) 255e266264
Author's personal copy
riparian areas to 20 m to aid native biodiversity conservation, thencalculated costs to their enterprises included reducing their herdsby 1.1e3.6% or increasing their purchase of fodder by between 4and 12%. Community contribution to farmer expenses could needto be considered to support farmer action for largely public benefit.A more detailed modelling exercise to inform policy developmentis required that includes weed control and on-going managementcosts and accounts for the lower relative DM production in riparianareas in non-drought and wet years.
Acknowledgements
The authors would like to thank the farmers who kindly assistedwith this research, allowing us access to their farms for pasturecollection and willingly supplying the required farm productiondata. We would also like to thank Ivor Awty for assistance withdeveloping the surveys. Murray Hannah contributed greatly to dataanalysis. Dan Armstrong provided advice and details for estimatingfarm productivity impacts. The authors would like to thank BillMalcolm for comments on an early draft, as well as anonymousreviewers for their helpful suggestions which improved thismanuscript. This project (MIS 06829) was funded by the VictorianDepartment of Environment and Primary Industries.
References
Aarons, S.R., 2011. Dairy farm impacts of fencing riparian land: an analysis offarmers’ perceptions of the costs and benefits. J. Soil Water Conserv. 66,140Ae147A.
Aarons, S.R., Gourley, C.J.P., 2012. The role of riparian buffer management inreducing off-site impacts from grazed dairy systems. Renew. Agric. Food Syst.28, 1e16.
Beedell, J.D.C., Rehman, T., 1999. Explaining farmers’ conservation behaviour: whydo farmers behave the way they do? J. Environ. Manag. 57, 165e176.
BOM, 2012. Australian Government Bureau of Meteorology e Climate Data On-line. Available from: http://www.bom.gov.au/climate/data/ (accessed 14.06.12.).
Boxelaar, L., Paine, M., 2005. The Social Dimensions of On-farm Change to ImproveWater Quality and Biodiversity. Report prepared for The University of Mel-bourne, Parkville.
Castro-Hermida, J.A., García-Presedo, I., Almeida, A., González-Warleta, M., CorreiaDa Costa, J.M., Mezo, M., 2009. Detection of Cryptosporidium spp. and Giardiaduodenalis in surface water: a health risk for humans and animals. Water Res.43, 4133e4142.
Cocklin, C., Dibden, J., Mautner, N., 2006. From market to multifunctionality? Landstewardship in Australia. Geogr. J. 172, 197e205.
Colwell, J.D., 1963. The estimation of the phosphorus fertiliser requirements ofwheat in southern New South Wales by soil analysis. Aust. J. Exp. Agric. Anim.Husb. 3, 190e198.
Cunningham, P.J., Blumenthal, M.J., Anderson, M.W., Prakash, K.S., Leonforte, A.,1994. Perennial ryegrass improvement in Australia. N. Z. J. Agric. Res. 37, 295e310.
Curtis, A., De Lacy, T., 1996. Landcare in Australia: does it make a difference?J. Enviro. Manag. 46, 119e137.
Curtis, A., Robertson, A., 2003. Understanding landholder management of riverfrontages: the Goulburn Broken. Ecol. Manag. Restor. 4, 45e54.
Dabney, S.M., Moore, M.T., Locke, M.A., 2006. Integrated management of in-field,edge-of-field, and after-field buffers. J. Am. Water Resour. Assoc. 42, 15e24.
Dorioz, J.M.,Wang, D., Poulenard, J., Trévisan, D., 2006. The effect of grass buffer stripsonphosphorusdynamicsea critical reviewand synthesis as abasis for applicationin agricultural landscapes in France. Agric. Ecosyst. Environ. 117, 4e21.
Doyle, P.T., Stockdale, C.R., Lawson, A.R., Cohen, D.C., 2000. Pastures for Dairy Pro-duction in Victoria. Report prepared for Department of Natural Resources andEnvironment, The State of Victoria, pp. 15e16.
DPI, 2006. Productive Grazing, Healthy Rivers: Improving Riparian and In-streamBiodiversity, Project Report 2006. Report prepared for Department of PrimaryIndustries. Available from: http://www.dpi.vic.gov.au/dpi/vro/vrosite.nsf/pages/lwm_biodiversity_publications#pghr-final (accessed 14.06.12.).
Frimpong, E.A., Lee, J.G., Ross-Davis, A.L., 2007. Floodplain influence on the cost ofriparian buffers and implications for conservation programs. J. Soil WaterConserv. 62, 33.
Fulkerson, W.J., Donaghy, D.J., 2001. Plant-soluble carbohydrate reserves andsenescence e key criteria for developing an effective grazing managementsystem for ryegrass-based pastures: a review. Aust. J. Exp. Agric. 41, 261e275.
Gourley, C.J.P., Melland, A.R., Waller, R.A., Awty, I.M., Smith, A.P., Peverill, K.I.,Hannah, M.C., 2007. Making Better Fertiliser Decisions for Grazed Pastures inAustralia. Report prepared for Department of Primary Industries.
Greiner, R., Patterson, L., Miller, O., 2009. Motivations, risk perceptions and adoptionof conservation practices by farmers. Agric. Syst. 99, 86e104.
Hennessy, D., O’Donovan, M., French, P., Laidlaw, A.S., 2008. Factors influencingtissue turnover during winter in perennial ryegrass-dominated swards. GrassForage Sci. 63, 202e211.
Isbell, R., 2002. The Australian Soil Classification. CSIRO Publishing, Melbourne.Jeffrey, S.J., Carter, J.O., Moodie, K.B., Beswick, A.R., 2001. Using spatial interpolation
to construct a comprehensive archive of Australian climate data. Environ.Model. Softw. 16, 309e330.
Lankester, A., Valentine, P., Cottrell, A., 2009. ‘The sweeter country’: social di-mensions to riparian management in the Burdekin rangelands, Queensland.Aust. J. Environ. Manag. 16, 94e102.
Lazenby, A., 1988. The grass crop in perspective: selection, plant performanceand animal production. In: Jones, M.B., Lazenby, A. (Eds.), The Grass Crop ethe Physiological Basis of Production. Chapman and Hall Ltd, London, pp.320.
Lovell, S.T., Sullivan, W.C., 2006. Environmental benefits of conservation buffers inthe United States: evidence, promise, and open questions. Agric. Ecosyst. En-viron. 112, 249e260.
Lynch, L., Brown, C., 2000. Landowner decision making about riparian buffers.J. Agric. Appl. Econ. 32, 585e596.
Malanson, G.P., 1993. Riparian Landscapes. Press Syndicate of the University ofCambridge, Cambridge.
Marques da Silva, J., Peça, J., Serrano, J., de Carvalho, M., Palma, P., 2008. Evaluationof spatial and temporal variability of pasture based on topography and thequality of the rainy season. Precis. Agric. 9, 209e229.
McKergow, L.A., Weaver, D.M., Prosser, I.P., Grayson, R.B., Reed, A.E.G., 2003. Beforeand after riparian management: sediment and nutrient exports from a smallagricultural catchment, Western Australia. J. Hydrol. 270, 253e272.
Melbourne Water, 2012. Melbourne Water Stream Frontage Management Pro-gramme. Available from: http://www.melbournewater.com.au/getinvolved/applyforfunding/Pages/Stream-frontage-management-program.aspx (accessed22.09.13.).
Moir, J.L., Scotter, D.R., Hedley, M.J., Mackay, A.D., 2000. A climate-driven, soilfertility dependent, pasture production model. N. Z. J. Agric. Res. 43, 491e500.
Moot, D., Mills, A., Lucas, D., Scott, W., 2009. Country Pasture/Forage ResourceProfiles. Report prepared for Food and Agriculture Organisation of the UnitedNations, pp. 38e39.
Naiman, R.J., Décamps, H., 1997. The ecology of interfaces: riparian zones. Annu. Rev.Ecol. Syst. 28, 621e658.
Nanere, M., Fraser, I., Quazi, A., D’Souza, C., 2007. Environmentally adjustedproductivity measurement: an Australian case study. J. Environ. Manag. 85,350e362.
Olsen, S.R., Cole, C.V., Watanabe, F.S., Dean, L.A., 1954. Estimation of AvailablePhosphorus in Soils by Extraction with Sodium Bicarbonate. USDA Circular. 939.
Pannell, D.J., Marshall, G.R., Barr, N., Curtis, A., Vanclay, F., Wilkinson, R., 2006.Understanding and promoting adoption of conservation practices by rurallandholders. Aust. J. Exp. Agric. 46, 1407e1424.
Parminter, T., Nelson, T., 2003. Dairy farmers’ evaluation of biodiversity prac-tices in south west Victoria. In: Proceedings of the Extending Extension:Beyond Traditional Boundaries, Methods and Ways of Thinking.Australasia-Pacific Extension Network National Forum 26 to 28 November,Hobart, Tasmania.
Parminter, T., Pedersen, J., 2000. Riparian Management Survey. A Survey toQuantify the Use of Riparian Management Practices and Farmer’s Attitudestowards Water Quality Management. Report prepared for AgResearch;unpublished.
Patrick, I., Barclay, E., Reeve, I., 2009. If the price is right: farmer attitudes to pro-ducing environmental services. Aust. J. Environ. Manag. 16, 36e46.
Pearson, C.J., Ison, R.L., 1997. Agronomy of Grassland Systems, second ed. CambridgeUniversity Press, Cambridge.
Peverill, K.I., Briner, G.P., Walbran, W.I., 1974. Problems associated with soil testingfor sulphur. In: Proceedings of the Transactions of the 10th InternationalCongress of Soil Science 2 to 6 September, Moscow, Russia.
QCCCE, 2012. SILO Climate Data. Available from: http://www.longpaddock.qld.gov.au/silo/index.html (accessed 22.09.13.).
Rayment, G.E., Higginson, F.R., 1992. Australian Laboratory Handbook of Soil andWater Chemical Methods. Inkata Press, Melbourne.
Rhodes, H.M., Leland Jr., L.S., Niven, B.E., 2002. Farmers, streams, information, andmoney: does informing farmers about riparian management have any effect?Environ. Manag. 30, 665e677.
Stockdale, C.R., 1983. Irrigated pasture productivity and its variability in the Shep-parton region of northern Victoria. Aust. J. Exp. Agric. 23, 131e139.
Stockdale, C.R., 1999. The nutritive characteristics of herbage consumed by grazingdairy cows affect milk yield responses obtained from concentrate supplemen-tation. Aust. J. Exp. Agric. 39, 379e387.
Straker, A., Lowe, K., 2004. Native Biodiversity Resource Kit - Environmental Manage-ment in Agriculture. Department of Sustainability and Environment, Melbourne.
Sullivan, T., Moore, J., Thomas, D., Mallery, E., Snyder, K., Wustenberg, M.,Wustenberg, J., Mackey, S., Moore, D., 2007. Efficacy of vegetated buffers inpreventing transport of fecal coliform bacteria from pasturelands. Environ.Manag. 40, 958e965.
S.R. Aarons et al. / Journal of Environmental Management 130 (2013) 255e266 265
Author's personal copy
Vanclay, F., 2004. Social principles for agricultural extension to assist in the pro-motion of natural resource management. Aust. J. Exp. Agric. 44, 213e222.
Weaver, D.M., Reed, A.E.G., Grant, J., 2001. Relationship between stream order andmanagement priority: a water quality case study. In: Proceedings of the ThirdAustralian Stream Management Conference, 27 to 29 August, Brisbane,Queensland, Australia.
Westra, J.V., Zimmerman, J.K.H., Vondracek, B., 2005. Bioeconomic analysis ofselected conservation practices on soil erosion and freshwater fisheries. J. Am.Water Resour. Assoc. 41, 309e322.
Wilcock, R., Monaghan, R., Thorrold, B., Meredith, A., Betteridge, K., Duncan, M.,2007. Land-water interactions in five contrasting dairying catchments: issuesand solutions. Land Use Water Resour. Res. 7, 1e10.
S.R. Aarons et al. / Journal of Environmental Management 130 (2013) 255e266266
