View metadata, citation and similar papers at core.ac.uk brought to you by CORE Available online at www.science.canterbury.ac.nz/nznsprovided by UC Research Repository Agricultural intensification, declining stream health and angler use: a case example from a brown trout stream in Southland, New Zealand Cohen Stewart1*, Rasmus Gabrielsson2, Karen Shearer2 and Robin Holmes2 1Southland Fish and Game Council, 17 Eye Street, Appleby, Invercargill 9810, New Zealand 2Cawthron Institute, 98 Halifax Street East, Nelson 7010, New Zealand * Corresponding author: [email protected] (Received February 2019, revised and accepted June 2019) Abstract The Otapiri Stream brown trout Salmo( trutta) fishery (Southland, New Zealand) has undergone a dramatic decline over the past five decades. Over the same period, agriculture has intensified in the catchment, switching from low intensity sheep and beef grazing to dairy support and forage crop. We compared historic (1960’s) and recent (2016) benthic invertebrate and electric fishing data and demonstrate substantial deterioration of the invertebrate community and reduced trout abundance. Compared with the baseline data, Ephemeroptera, Plecoptera and Trichoptera (EPT) abundance declined by 84% in riffles. The greatest decline (98%) occurred in pool habitat. Disproportionately high decline of EPT abundance in pool habitat highlights a potential issue with standard invertebrate sampling methodologies used to collect data to inform national-scale stream health assessments. By sampling only faster water habitat types (i.e., riffle and runs), the impacts of land use intensification (e.g., sedimentation) could be being underestimated. In line with the observed decline in %EPT abundance, brown trout density has declined by approximately 71% and angler use of the fishery declined by 76% between 1994 and 2016. We contend that the decline in angler use of the Otapiri Stream is a result of land use intensification within the catchment, driving changes in the trout invertebrate food- base (EPT abundance) leading to reductions in adult trout density. Key words: Angling, brown trout, Salmo trutta, macroinvertebrates, water quality, land use intensification New Zealand Natural Sciences (2019) 44: © New Zealand Natural Sciences 2 New Zealand Natural Sciences 44 (2019) Introduction ciety Field Officer, recalls frequent angler use of the Otapiri when undertaking angler Intensive agriculture is increasing in extent compliance activities in the 1980’s. Expert in New Zealand, particularly in Southland Southland angler Len Prentice recalls the (Statistics NZ, 2017). In 1990, there were Otapiri Stream to be of exceptional fishing 37,772 dairy cows in Southland whilst quality during the 1970’s and 80’s, with the in 2016 there were 708,895 (Statistics Otapiri offering mayfly hatches through- NZ, 2017). There is now overwhelming out the day and excellent dry fly fishing. evidence that intensifying agricultural More recently (1994-2016), angler use of land use leads to degraded water and the Otapiri fishery has declined substan- stream habitat quality (Matthaei et al. tially (Unwin 2016), and local anglers have 2006; McDowell & Wilcock 2008). In regularly expressed concern about reduced New Zealand, stream invertebrates are trout abundance and loss of mayflies in the commonly used by regional councils to stream. Reduced recreational value of the monitor the health of streams. However, Otapiri Stream is concerning for fisheries the State of Environment Reporting managers because the Otapiri fishery is programme (SoE) which includes the valued by anglers for ease of access and its collection of stream invertebrate samples, proximity to Southland’s main population was not initiated in Southland until centres (Unwin 2013). To investigate how 1997. Prior to this period, agriculture was the Otapiri Stream invertebrate commu- intensifying in Southland (Statistics NZ, nity has changed over the last 60 years, we 2017). Consequently, there is little pre- resurveyed the invertebrate and brown intensification baseline invertebrate data trout population and compared these for many Southland streams. However, in survey results with historical Southland the early 1960’s the Marine Department Acclimatisation Society data. Our aim was and Southland Acclimatisation Society to observe whether substantial changes (now Southland Fish & Game), undertook in the macroinvertebrate communities benthic invertebrate sampling and electric had occurred in parallel with agricultural fishing surveys in the Otapiri Stream, a intensification within the catchment, and hill country stream that flows through if changes to the trout invertebrate food- the Southland Plains. These data provide base are a plausible explanation for the a valuable resource for investigating the observed decline in the quality and usage historical ecological state of a ‘typical’ of the Otapiri fishery. hill country/lowland stream prior to the conversion of much of the catchment Methods from low intensity sheep and beef to dairy support and forage crop land uses. Study site and survey timing From 1960 through to the 1980’s, both formal and anecdotal records show the The Otapiri Stream is a 5th order, rain- Otapiri Stream was highly valued amongst fed hill country stream with a stony bot- anglers. The Otapiri had the highest angler tom and natural flow regime. The Otapiri catch rate of any Southland waterbody stream has its origins in the Hokonui Hills and was identified as the most important and flows into the Makarewa River, a major Makarewa River tributary for anglers tributary of the Oreti River. The median (Graynoth & Skrzynski 1974). Mark Sut- Otapiri Stream flow is 1.08 m3/s and the ton, former Southland Acclimatisation So- catchment land use is predominantly STEWART ET AL.: DECLINING STREAM HEALTH AND ANGLER USE OF A TROUT STREAM 3 agricultural - sheep and dairy pasture and and preserved with ethanol. In the labora- forage crop with pockets of native podo- tory the invertebrate samples were sorted carp and exotic pine (Pinus sp.) forest. and identified to the lowest taxonomic The Otapiri catchment is situated in a hill level using the keys of Winterbourn et al. country/bedrock physiographic zone and (2006) and Chapman et al (2011).. Chi- features of this land type include: rolling/ ronomids were subsampled where abun- steep land, high rainfall and a dense net- dances were very high using a volumetric work of branching streams that flow to approach whereby a 500 ml invertebrate neighbouring lowland areas which are a sample was evenly mixed then 100 ml was potential source of nutrients and sediment removed for processing. (Environment Southland, 2017). Invertebrate health metrics calculated In 1963-64, Boud & Cudby (1966) col- included: the Macroinvertebrate Commu- lected benthic invertebrate samples from nity Index (MCI), Quantitative Macroin- five distinct sites in the Otapiri Stream, vertebrate Community Index (QMCI) ranging from the upper reaches (above the (Stark & Maxted 2007), %EPT (by taxa Otapiri Gorge) to the lowermost reaches and abundance) and invertebrate density. (near the Makarewa River confluence). Brown trout density At each site invertebrate samples were collected from pools (n = 9) and riffles (n To survey the brown trout population at = 9)in February, May and August of 1963 the Otapiri Gorge site, Boud & Cudby and January 1964. Average invertebrate (1966) undertook summer (January species abundance over the four sampling 1964) electric fishing surveys (pack-set periods was determined for each site and machine). Boud & Cudby (1966) did habitat type (riffle/pool). For the current not specify the number of electric fishing study, we sampled benthic invertebrates at passes performed. In February 2012, we a single site during February 2016, in the surveyed the brown trout population of lower Otapiri Gorge area. This site was a single reach (pool, riffle, run) using a chosen because it is of most interest to three-pass electric fishing survey (Temple anglers and was also a survey location used and Pearsons 2007) and determined brown by Boud & Cudby (1966) (46° 2’3.18”S, trout density in the selected Otapiri Stream 168°26’47.49”E). reach (46° 2’42.21”S, 168°26’50.37”E). Our electric fishing survey was conducted Benthic invertebrate sampling as part of an earlier study and was not the At the lower Otapiri Gorge site, we reach surveyed by Boud & Cudby (1966). selected one pool and one riffle to conduct However, our electric fishing survey site benthic invertebrate surveys. Six benthic was within 150m of the site used by Boud invertebrate samples were collected using a & Cudby (1966) and involved electric Surber sampler (0.25 m2, 0.5 mm mesh) at fishing in a comparable section of river each habitat type (i.e., pool and riffle) (see which included pool, riffle and run habitat. Stark et al. 2001, Protocol C3). A transect Angler use line was placed along the length and width of the pool and riffle, and sampling We used the Fish and Game National locations selected using a random number Angler Survey (Unwin 2016) to assess how generator. angler use of the Otapiri has changed over Following collection, the invertebrate time. Unwin (2016) provides information samples were placed in labelled containers on angler usage of New Zealand fisheries 4 New Zealand Natural Sciences 44 (2019) between 1994-2016. 1). However, the invertebrate commu- nity in the 2016 survey was dominated Results by Diptera (chironomids) in the riffles and Mollusca (Potamopyrgus sp.) in the Benthic Invertebrates pools (Figure 1). In 1963-64, EPT taxa Boud & Cudby (1966) did not include (by abundance) made up 60% and 55% of raw invertebrate data in their report but the invertebrate population in the riffles did provide the average abundance for each and pools respectively (Boud & Cudby invertebrate species over the four sampling 1966). In 2016, EPT taxa were only 9.4% periods for each habitat type (pool and and 1.3% of the invertebrate population in riffle). As such, we could not statistically riffles and pools respectively (Table 1). In compare the differences in stream health most cases, the abundance of EPT inverte- metrics by sampling period.