Part 2 - Understanding your Analogue Ecosystem
In this series of articles Katina Strelein, Astron’s Principal Scientist – Rehabilitation, discusses the importance of being able to cost effectively collect representative data over a large area as both the basis for setting completion criteria and for monitoring and assessing performance against them (View Part 1 - What are Completion Criteria). This article focuses on why it is critically important to have a thorough understanding of the reference or analogue ecosystem if comparison with a similar ecosystem forms the basis for the assessment of rehabilitation performance.
Basic to the selection and use of reference or analogue information is the need to understand temporal and spatial variation in ecosystems (White and Walker 1997). But why is the selection and use of reference information difficult? Because nature is variable in time and space and we often must use fragmentary information about this variation to understand current conditions, deduce the potential for rehabilitation, set completion criteria and propose rehabilitation methods (White and Walker 1997).
Assessment of analogue and rehabilitation sites has conventionally been conducted using on ground monitoring of quadrats or transects. Effective on ground sampling is possible, but it requires good design, technical expertise and often time and money. However, these resources are not always sufficiently available. To be successful a monitoring program needs to:
- Achieve sampling adequacy and therefore statistical power (http://www.astron.com.au/news/environmental-monitoring-designs/). For example, if monitoring is not conducted across seasons and years, temporal variation, including the dynamics produced by disturbance and succession, may not be recorded. If analogue sites are only monitored in a particularly favourable season this can set an unrealistically high bar for rehabilitation to meet.
- Use quantitative techniques or mitigate the problems associated with estimation. Consistent recording of data in the field requires experienced personnel and cross calibration. For example, in a study on the repeatability of cover estimates Cheal (2008) found that 16 people with technical or tertiary biological qualifications (including experience in standard field survey) rated the projected foliar cover of the same area of Triodia differently, with scores ranging from 10% to 60%.
- Be based on good scientific design which includes:
a). Randomisation - to remove the bias that can be introduced by placement of transects or quadrats. The results of Peck et al (2015) support the need to draw on multiple reference sites and metrics to account for the variability associated with naturally dynamic ecosystems.
b). Stratification if community or ecotype are clearly present.
Due to the cost of conducting field work, commonly only 2 or 3 transects or quadrats are established and monitored within each rehabilitation or analogue area. For Landscape Function Analysis (LFA) monitoring, it is recommended that two transects are sufficient per area (Tongway and Hindley 2005). Due to the heterogeneity of vegetation, this limited number of transects/quadrats may not provide data that is representative of the whole area that is being assessed.
Regional variation is often such that no one reference site is ever a perfect match for a rehabilitated area; rather, our conceptual model should be one of interpolation among multiple sites and sources of information (White and Walker 1997). Similarly, ecosystems are likely to have unique histories so that no one reference sites observed at an arbitrary time should be used to determine goals for rehabilitation (Pickett and Parker 1994).
Currently a lack of representative data over large areas limits our ability to adequately understand analogue or reference ecosystems and set appropriate completion criteria. The rapidly advancing emerging field of remote sensing is one methodology that allows monitoring to be undertaken cost effectively at the landscape scale so that the performance of a whole landform and the vegetation surrounding it can be understood. The final article in this series provides more detailed information on landscape scale monitoring and how it can be used for rehabilitation monitoring.
However, even with improved data collection technologies and cheaper ways to assess more of the landscape – is comparison with analogue or reference transects the best way to assess rehabilitation success? In the next article (Part 3) some potential alternative methods for the assessment of rehabilitation performance are put forward for consideration.
Cheal D 2008, ‘Repeatability of cover estimates?’, Ecological Management and Restoration, vol. 9(1), pp. 67-68.
Peck JLE, Commarmot B, Hob ML and Zenner EK 2015, ‘Should reference conditions be drawn from a single 10 ha plot? Assessing representativeness in a 10,000 ha old-growth European beech forest’, Restoration Ecology, vol 23(6), pp. 927-935.
Pickett STA and Parker VT 1994, ‘Avoiding old pitfalls: opportunities in a new discipline’, Restoration Ecology, vol 2, pp. 75-79.
Tongway DJ and Hindley NL 2005, Landscape Function Analysis: Procedures for Monitoring and Assessing Landscapes With special reference to Minesites and Rangelands.
White PS and Walker 1997, ‘Approximating Naure’s Variation: Selecting and Using Reference Information in Restoration Ecology’, Restoration Ecology, vol. 5(4), pp. 338-349.