Geoscience Reference
In-Depth Information
This chapter describes the development of a
systematic approach to biodiversity conservation
planning for New Zealand's rivers and streams, and
its use is demonstrated through the identification
of a representative set of high value sites for
management as part of a wider application of
conservation planning tools within this country's
Department of Conservation. Specific aspects that
are described include the development of a
biologically optimized environmental classification
for use as a surrogate of biodiversity patterns,
expert assessments of human impacts for use
as a surrogate of biodiversity condition, and the
complementarity-based estimation of site priority
using an approach that incorporates explicit
consideration of longitudinal connectivity.
relationships between biological turnover and a set
of functionally based environmental predictors at
sample sites for freshwater fish (13 363 sites) and
macroinvertebrates (2677 sites); results were then
combined to guide the subsequent classification
process. The classification contains four levels of
classification detail (recognizing 20, 100, 200 and
300 river groups respectively) to provide differing
levels of discrimination for different applications.
At a broad (20-group) level of detail (Table 15.1)
groups A-F occur predominantly in the lowlands,
with groups A (lowland, low-gradient streams)
and C (lowland hill-country streams) the most
extensive; groups G-L occur at middle elevations,
varying in their frequency of high-flow events and
degree of glacial influence; group M contains a
very small set of larger, glacially fed lowland rivers;
and groups N-T occur at high elevations differing
mostly in their elevation and degree of glacial
influence.
Estimates of biodiversity condition were based
on the systematic assessment of human pressure
factors, measured across all river segments using
a mix of GIS-based and modelled data (Abell
et al
., 2002; Leathwick
et al
., 2010). Estimates
of catchment clearance and impervious cover
were derived from GIS analysis of satellite
and digital topographic map data, respectively.
Nitrogen inputs were estimated from a regionally
based regression model implemented within a
catchment framework (Woods
et al
., 2006). The
distributions of introduced fish were predicted
from environment-based statistical models using
catch data for nine species from an extensive
collection of fish sampling data (n
Methods
Input data
The analysis was carried out in three stages
involving the preparation of input data, the
calculation of three sets of site priorities, and
the selection of high value sites. Input data
were derived from a digital description of New
Zealand's river network derived by GIS analysis of
digital elevation data (Snelder and Biggs, 2002).
This network contains approximately 567 000 line
segments, each representing a section of river
or stream between two adjacent confluences. All
analyses used planning units (Plate 18a) that break
rivers of fourth or higher order into their third or
higher order sub-catchments and their main stem.
Catchments of first- to third-order streams were
treated as individual planning units.
Attribute data associated with each river and
stream segment described both its membership in
an environmental classification (Plate 18a) and
the estimated current biodiversity condition given
human impacts (Plate 18b). The environmental
classification, described in detail by Leathwick
et al
. (2011), was derived using generalized
dissimilarity modelling (GDM - Ferrier
et al
.,
2007) to maximize its ability to capture biological
patterns. Separate GDM analyses were made of
13 363).
The locations of mines, dams and other point
discharges were derived from digital spatial data
from a range of sources, with impacts modelled
upstream and/or downstream as appropriate.
Expert opinion was used to define a response curve
describing the expected impact of each pressure
on biodiversity, with the estimated impacts for all
factors combined to give an overall estimate of
pressure. Biodiversity condition was then defined
as 1minus total pressure, values ranging from zero
(totally degraded) to 1 (pristine) (Plate 18b).
=
