Stream Ecosystem Responses to Retention, or Restoration and Maturation of Urban Forests in Comparative New Zealand Watersheds.

Abstract

In New Zealand the city-edge land use transition from rural to urban provides an opportunity to rehabilitate riparian forests and streams as part of Low Impact Urban Design and Development (LIUDD; van Roon and van Roon, 2009). The protection of ecosystem services and downstream water quality, the buffering of climate change impacts on watersheds, and the halting of biodiversity loss necessitate the protection and restoration of riparian terrestrial and aquatic ecosystems. Conditions need to support permanent riparian forest cover within the urban matrix. The quality and long-term viability of restored ecosystems needs to be ensured if the objectives above are to be achieved. This paper reports on paired watershed studies over a 7-year period in Manukau, Auckland. These studies have enabled comparisons to be made between conventional and LIUDD subdivision at low (2000 – 5000 m2 lots) and average (400 - 500 m2 lots) residential densities. Factors used in these comparisons included the proportion of the watershed forested versus impervious, proximity of forest vegetation to streams, changes in forest structure and maturity with age, the degree of piping of streams, the presence and type of stormwater treatment devices, and the health of in-stream macro-invertebrate communities as indicators of stream and watershed condition. Results for in-stream health indicators are compared with values for the greater Auckland region and demonstrate the superiority of LIUDD forested neighbourhoods. Low density residential watershed comparison: Between 2005 and 2012 macro-invertebrate indices were intermittently monitored in the streams of 6 neighbouring watersheds where low-density residential development was either present and stable (2 control watersheds) or on-going (4 treatment watersheds). Prior to this time (2004) subdivision earthworks were carried out on part of the 4 treatment watersheds excluding stream gullies and steep faces. Scattered mature riparian trees were retained in valley floors. Subdivision streets and lots were laid out and drainage networks completed. On average, 60% of each of these treatment watersheds including stream riparian zones was planted in indigenous forest plants in 2004, the year prior to monitoring commencing. These planted areas receive dispersed and raingarden-treated stormwater from individual properties and streets. House construction has been progressing on the other 40% of each treatment watershed and approximately half of the 66 houses planned across the 4 watersheds have now been constructed. One of the control watersheds has more forest than the other. Average density residential watershed comparison: Of the 4 watersheds in this comparison 3 are treatment watersheds and 1 is a control. The control watershed contains no urban forest, the stream is piped and there is no stormwater treatment. The 3 treatment watersheds have had on-going subdivision development prior to and during the monitoring period, and all have retained and protected mature riparian forest. Stormwater treatment and discharge to streams varies between these 3 watersheds. Stream monitoring shows that the support given by forests or re-establishing plant communities in riparian corridors to healthy stream ecosystems is demonstrable but it may be undermined by discharges from some types of stormwater or wastewater discharges.