WALPOLE FINE GRAIN FIRE MOSAIC
(WFM)
PROJECT BACKGROUND
AND
INVERTEBRATE METHODS
Dec 2004
Creating a Small Grain Mosaic with Frequent Fire: Long Term Impacts
on Biodiversity.
Executive Summary
This is a novel and exciting collaborative operational trial involving staff from the
Department’s Science Division and the Frankland District to a) determine whether a
small-grained fire-induced mosaic can be created by the frequent introduction of fire into
the landscape and b) to investigate the impacts of this mosaic on elements of the
biodiversity. There is ample scope for involving volunteers and community-based groups
or individuals in this project, which is part of an adaptive management approach to fire
management in the Warren Region. The project is large-scale (several thousand ha) and
the sampling methodology is based on that used for FORESTCHECK, which was
developed out of extensive consultation with scientists external to the Department. In this
context, the monitoring sites have been re-named Walpole Fine Grained Mosaic sites.
The proposal will be submitted for partial funding through the Bushfire CRC as part of
that CRC’s Program B.1.
Introduction
Flammable vegetation and hot, dry summers have ensured that fire is a natural environmental
factor, which together with climate, landform and soils, has operated over thousands of years to
forge the remarkable biodiversity of south-west ecosystems. Plants, animals and ecosystems have
evolved in this fire prone-environment and have developed a range of physical and behavioural
traits that enable them to persist with, and in some cases, depend upon a variety of fire regimes
For many plant species, reproduction and regeneration are cued or enhanced by fire and for many
plant communities, particular fire regimes are necessary for the maintenance of floristic and
structural diversity. A particular sequence and scale of fires are necessary to provide habitat
diversity and opportunity for animals. However, the way in which species and communities
respond to fire is variable. Some assemblages are quite resilient to frequent fire and recover to
their pre-fire state relatively quickly, while others are more sensitive to fire and can take many
decades to recover. Thus, no fire regime, or history of fire interval, season, intensity, patchiness
and scale is optimal for all species and communities and although fire diversity can promote
biodiversity, some fire regimes can also threaten biodiversity.
The damage potential and difficulty of suppression of a bushfire is determined by the amount of
vegetation that burns, its moisture content and the weather conditions. Prior to European
settlement, burning by Noongar people probably maintained a patchwork-quilt of vegetation at
different stages of post-fire development (pyric stages) – from recently burnt patches to long
unburnt patches, although there is mounting evidence that much of the landscape was maintained
in an early post-fire state. This fire mosaic contained the spread and intensity of wildfires. Very
large and intense wildfires were not in the best interests of Aboriginal people, nor the
environment generally, and were probably rare events.
On lands managed by CALM, the primary aim is to manage fire to conserve biodiversity and to
ensure an acceptable level of protection to human life, property and industry. Fire management is
complex and potentially dangerous and requires the skilful combination of art and science. Fire
science, including fire behaviour and ecological effects of fire on natural ecosystems, has
advanced as a result of on-going research undertaken by a range of organizations over the last 40
years or so. Although our knowledge is incomplete, fire management must be underpinned by
science.
With respect to the ecological effects of fire, several themes consistently emerge from the
contemporary scientific literature. Firstly, no one fire regime suits all organisms. For example,
some elements of the biota benefit from frequent fire (e.g., grasses, annual herbs, some resprouters, large macropods, some fungi and some invertebrate groups), while others are
disadvantaged (e.g., some late maturing obligate seeders, some mammals, some birds). At the
other end of the spectrum some elements benefit from longer intervals between fire and others are
disadvantaged. A second emergent theme for conserving biodiversity at the landscape scale is
that of spatial heterogeneity, or of a small-grained fire-induced habitat mosaic. A third theme to
emerge is the need to protect relatively fire sensitive organisms and ecosystems such as riverine
or riparian zones, aquatic ecosystems, some wetland and heathland complexes and granite
outcrops from frequent fire or large and intense wildfire.
Creating fire-induced mosaics
Hence, the fire management challenge in many south-west landscapes is to devise and implement
fire management strategies that are diverse, provide a fine-grained fire-induced mosaic, that
protect fire sensitive elements, that enable the periodic regeneration of fire-promoted elements
and habitats, that minimises the impact of wildfires and that is practical and affordable.
There are a number of approaches to achieving these objectives and they include:

Vary the fire regime applied to a fire management unit. This includes varying the season,
frequency and interval of fire based on vital attributes and life histories of key taxa (fire
sensitive taxa).

Implement mostly patchy burns with occasional near complete burn. That is, create
patchiness within patches (an intra-patch mosaic).

Small-grained fire-induced mosaics can be achieved several ways including (but not limited
to):

Physically breaking the management unit up with a network of mineral earth or natural
physical fire barriers and treating each cell differently. This is generally unacceptable and
undesirable.

Introducing fire into the landscape when moisture differentials exist across the landscape
during winter and spring particularly. This comes with the risk of re-ignition and
complete burn-out as the season and vegetation dries with the onset of summer.

.Using wind-driven strips in heath or mallee heath vegetation.

Lighting/ignition pattern and time of ignition.

Introducing fire into the landscape frequently when fuel flammability (quantity and
cover) differentials exist (similar to the likely Noongar Aboriginal burning regime). The
resultant mosaic and effects on biodiversity at the landscape scale over time is unknown
but should be trailed and monitored.
Trial Objectives
The techniques described above (and there are probably others) alone or in combination can be
used to create a small-grained mosaic over time. Each technique has advantages and
disadvantages and each will result in a different mosaic. Prescribed burning at 5-10 year intervals
during spring when moisture differentials exist has been routinely applied over many areas for
several decades. However, there is the risk of re-ignition later, leading to burning out of the
entire landscape and to possible fire escape.
Alternative techniques, such as frequently introducing fire into the landscape in summer/autumn
has not been trialed and properly monitored on a large scale (thousands of hectares) anywhere.
Throughout the south-west forest region (Jarrah Forest and Warren Bioregions), the rate of postfire production of live and dead vegetation, which becomes the fuel for a bushfire, is such that it
is insufficiently mature as a fuel until about 3-4 years after fire. Prior to this, fire spread, even
under conditions of high fire danger, is likely to be very patchy and restricted to relatively small
areas with a dense overstorey canopy cover. However, beyond about 3-4 years after fire, more of
the landscape is flammable, and by about 5-8 years after fire, the vegetation/fuel complexes are
sufficiently well developed so that virtually the entire landscape will support an intense bushfire.
Fires burning in light and patchy (immature) fuels will not only be very patchy, but will be low
intensity. This combination of very patchy fires burning at very low intensity and only burning in
the more flammable parts of the landscape provides opportunity for fire sensitive species and
communities, that mostly occur in less flammable parts, to either not carry fire, or be protected
from fire by intervening low fuel patches, or survive low intensity fire. We hypothesise that over
time, the frequent introduction of fire into the landscape every 1-3 years will result in a finegrained mosaic of patches at different stages of post-fire development. The precise nature and
scale of the mosaic is unknown. In addition, we hypothesise that the resultant mosaic will afford
protection to fire sensitive species and communities and will significantly break up the run and
intensity of an impinging a wildfire. It is stressed that because of the patchy nature of fuels, the
frequent introduction of fire does not equate to frequent burning. We hypothesise that a relatively
small proportion of the landscape will in fact carry fire, hence burn, each year.
We propose the implementation of a carefully monitored landscape-scale trial to investigate the
long term effects of the frequent introduction of fire (1-3 years) on:

The scale, and heterogeneity and spatial and temporal dynamics of the resultant fire-age
mosaic.

Abundance of fire sensitive plant taxa, DRF and reserve listed taxa (e.g., Synaphaea sp.,
Lambertia rariflora subsp. lutea, Banksia quercifolia, B. occidentalis, Dryandra
formosa, Hakea oliefolia)

Geophytes (Orchidaceae, Droseraceae especially)

Abundance of fire sensitive fauna and condition of associated habitats (e.g., Honey
Possum (Tarsipes rostratus), Quokka (Setonix brachyurus), Mardo (Antechinus
flavipes), Sunset Frog (Spicospina flammocaerulea)).

Reptile assemblages

Bird assemblages

Invertebrate assemblages

Vascular plant assemblages

Fungi and cryptogams.
Specific taxa and habitats to be studied will depend on the selection of the study site.
Methods
Location
Prosed areas for this trial are a) the southern portion of Surprise block and b) London block,
Frankland District. Adjacent and matched sites will be selected as reference or control areas.
Reference areas will be subjected to the standard fire management treatments prescribed in the
various planning documents for the region.
The trial will be large scale (2,000 – 5,000 ha – site to be determined). It will be of the BACI
design (Before, After, Impact Control) with replications through time. Sampling will be stratified
according to 3-4 major vegetation types; e.g., Open Forest, Low Open Woodland, Sedgeland,
Myrtaceous Heath, Riverine. Final classification will depend on study site selection. Additional
sample site will be located where DRF, reserve listed, threatened fauna or fire sensitive taxa
occur.
The sampling protocols to achieve the objectives above will essentially follow those developed
for the FORESTCHECK monitoring program and for the purpose of this trial are termed Walpole
Fire Mosaic sites or grids. In this trial, all grids described below will be sampled at 3 intervals
initially then reviewed. A Nearby weather station (rain gauge) will need to be identified or
established if none exists (Note: max and min temps would also be useful for invertebrates,
especially if a complete data set of uninterrupted daily temperatures throughout the study were
available. This is not essential however).
Introducing Fire
The intention is to introduce, (but not burn out) fire into the study site. Donna/Greg/Karlene –
can you write something here about how we might go about this from an operational perspective?
Proposed Prescription (Suggestions?)
SDI
Jarrah
SMC
>1000
6-10%
Tower
wind
speed
8-15
kph
Jarrah
ROSI
Temp.
RH
Lighting
pattern
40-80
25-32
30-45
200m spacing?
(Comments)
Mapping the Mosaic
Recently developed satellite imagery technology will be used to map the fire-induced mosaic.
Imagery will be analysed and maps prepared after each introduction of fire into the study area.
Areas that did not burnt, areas that burnt and fire intensity will be mapped. A chronological
sequence of such maps will result in a map of the scale and nature of the mosaic (fire history for
various ecosystem elements or habitat types represented in the landscape). Identification and
mapping of ecosystem elements or habitat types will occur on final selection of the trial site. A
map of the existing mosaic on the study site will need to be prepared form historical imagery
captured after the most recent fires.
Broad vegetation assemblages/habitat types will be identified and mapped, as will locations of
DRF, reserve listed species, threatened fauna, fire sensitive taxa and communities. A fauna and
flora species list will be prepared for the study area (source: WA Museum, WA Herbarium, local
knowledge). Fire history (to 1937) is being prepared for the proposed Walpole Wilderness Area.
For grid lay out see example of Forestcheck grid. Below is basic plan of the Bird
census grid from Forest check
100 m
100 m
Invertebrates
Leader
Janet Farr
Members
Allan Wills, Paul VanHeurck, Karlene Bain
Objectives
WALPOLE FIRE MOSAIC (WFM) invertebrate monitoring is intended to provide a method to easily
capture data describing the abundance and community composition of readily visible and highly
identifiable components of the invertebrate fauna. Sampling methods are designed to employ
efficient use of available time and resources, effective coverage of habitats and thus potential
diversity, and minimal specimen processing time. The data are intended to provide a means for
monitoring the assemblage of invertebrate fauna subject to disturbances such as fire and natural
variation and will be directly comparable to data collected for FORESTCHECK. Sampling will
include ground, understorey and aerial habitats. In addition, Gondwanan relic invertebrates will
be noted in all habitats when observed (see FORESTCHECK for summary of Gondwanan taxa
and indicator species).
Sampling methods
Invertebrate monitoring will be conducted within the sampling grids established for flora,
vertebrates and birds as above. Timing of sampling is at present subject to discussion. From
experiences with Forestcheck it would seem a single annual sample in late spring may be
sufficient, however since this study is specifically examining fire effects, an autumn sample
would be useful in capturing species such as the “splendid ghost moths” of the genus Aenetus.
Two species of which (A. dulcis and A. scriptus) are exclusive to the south-west of WA and fly
from February to May. If we sample in spring (and autumn?) as for Forestcheck all sampling
strategies (see below) could be achieved except perhaps for light trapping. It is envisioned that a
conflict in use of light traps could occur with Forestcheck plots located in the southern forest
regions due to the narrow window of opportunity for this sampling method. However this could
be alleviated by either (1) disregarding the moon illumination requirements applied to
Forestcheck in WFM sites and/or (2) light trapping later in the season compared to the other
sampling conducted for the WFM project. Despite the logistical problems associated with light
trapping for this project, this sampling procedure is highly recommended for two main reasons
(1) from experiences with Forestcheck it is by far the most efficient sampling method; (2) this
method has not been applied systematically in the Walpole region before and, due to the
interesting botanical features of the intended study area, could capture some worthwhile insect
specimens. In addition preferred sampling times could clash with burning times therefore this
will need to be considered more thoroughly when burn prescriptions etc are known more fully.
Five types of sampling method will be undertaken which will involve both active and passive
(trapping) capture techniques:
1.
2.
3.
4.
5.
Pitfall trapping (passive).
Area searches of ground habitats including litter and coarse woody debris (active).
Canopy and bole observations of known forest pest species where host plants are present.
Foliage beating and sweep netting for day active Gondwanan orders and other indicator
species (active).
Light trapping for night flying Gondwanan orders and other indicator species (passive).
Because sample recoveries from all methods are likely to be affected by temperature and rainfall
conditions, air and soil temperatures and rainfall events will be recorded as additional covariates
during field sampling.
Light trapping
One trap per grid site to be placed in a clear area within the site (usually at or near grid centre) so
as to maximize light field attraction within the plot. The preferable killing mechanism is pest
strips (placed in the capture bucket) to maximize quality of specimens. Traps will be operated
nightly from sunset to sunrise with clearance the following morning. This will be repeated 3
times within each trapping season (3 wks), such that there will effectively be 3 trap nights per
quadrat grid for each sampling season (this is current protocol for FORESCHECK and may be
reduced). Minimum night temperatures will be recorded for each trap night at each site. Captures
will be placed in a paper bag for each site and clearly labelled with site details and date. Should
storage over 24 hrs be required, the paper bag containing specimens will be placed in a plastic
bag and the sample frozen. Trapping periods are preferably close to a new moon to avoid light
interference from a full moon (this is current protocol for FORESCHECK and moon illumination
may have to be disregarded in order to schedule sampling).
Foliage beating and net sweeping
Understorey canopy beating will be conducted for 1 hr using a beating tray (approx 840 x 840
mm). Net sweeping (using a 380mm standard hoop, a 121 cm telescopic handle and funnel
tapered net bag) will also be conducted over 1hr. Sampling times can be broken into ½ hourly
segments and each sampling technique alternated (e.g. ½ hour beating, ½ hour sweeping, ½ hour
beating, ½ hour sweeping). Net sweeping is not recommended during early mornings, as air
temperature may be too cool for insect activity. Thus adjustment in individual plot sampling
schedules may need to be made to accommodate insect activity variations. Collections are placed
in vials labeled according to site, capture method and date, and stored in a cool insulated
container (to prevent moisture and heat spoilage) on site. Bulked samples are frozen pending
laboratory sorting. Soft bodied specimens will need to be directly stored in 70% ETOH on site or
within 8 hrs to facilitate later identification. Soil and air temperatures will be taken at each site
during sampling and recorded at conclusion of sampling period.
Observations of known pest species
Three known pest species will be monitored (conditional on presence of host species).
1.
2.
3.
Jarrah Leaf Miner (JLM is a Gondwanan relic species). Jarrah canopy will be visually
scanned for 10 min using binoculars. Evidence of JLM presence (e.g. shot holes in leaves)
will be recorded as 0 = absent, 1 = present, 2 = abundant. Abundant JLM will be classed as
greater than or equal to 5 shot holes per leaf present.
Gum Leaf Skeletonizer (GLS). Both the upper canopy of jarrah, marri, Corymbia ficifolia
and other foliage hosts, and understorey will be visually scanned (10 min) for (a) leaf
damage peculiar to GLS, (b) GLS egg rafts, (c) GLS moulting casts, (d) GLS larvae. Each
individual symptom will be scored as present or absent. In addition a general score for
abundance level will be given as 0 = absent, 1 = present, 2 = abundant (greater or equal to 5
symptom groups found). Some locations around Walpole have formerly been found to
support outbreaking populations and a 2 generation per year population of GLS, thus if
actual specimens of the larvae are found, their life stage will be recorded
Bullseye Borer. Tree boles of all susceptible hosts will be scanned for borer symptoms (10
min) (e.g. kino dribble stains, bulls eyes, helical scribing in bark). Recording will be 0 =
absent, 1 = present, 2 = abundant. Abundant will be classed as greater than 4 symptoms per
tree.
Area searches within microhabitats
Within each quadrat, sampling by area searches will be stratified according to at least two
microhabitat types.
1.
Leaf litter. Within litter beds. This category will include the searching of moss swards and
lower tree boles, underbark shelters and accessible hollows in lower boles below a height of
50cm.
2.
3.
On/in/under coarse woody debris.
Ashbeds. Areas of ground upon which large pieces of coarse woody debris fallen logs and
branches have burnt completely or almost completely to ash may be included as an
additional habitat when abundant but otherwise will be included in coarse woody debris
Sampling will be conducted for 1 hr in each habitat (eg 1 hr for litter searching and 1 hr for coarse
woody debris). This time may be divided between operators, such that if 3 operators are available
they will search for 20 min each. However, an individual operator search time of less than 20min
(ie. more than 3 operators) is not recommended since this may reduce sampling effectiveness.
Collections are placed in vials labeled according to site, habitat and date, and stored in a cool
insulated container (to prevent moisture and heat spoilage) on site. Bulked samples are frozen
pending laboratory sorting. Soft bodied specimens will need to be directly stored in 70% ETOH
on site or within 8 hrs of capture to facilitate later identification. Soil and air temperatures will be
taken at each site during sampling and recorded at conclusion of the sampling period.
Pitfall trapping
Ten traps at 10m intervals will be placed along the quadrat diagonal. Trap cup size is 90mm
diameter by 110 mm deep. Ethylene glycol is used as a trap preservative. Pitfall traps are open for
a ten-day period at all sampling sites simultaneously (due to the number of sample grids, grids
may be grouped for collection convenience). Pitfall traps will be open for 3 sets of 10 days for
each sampling season. At trap closure combine the contents of traps 1 to 5 in one sample and
traps 6-10 as a separate sample (this is protocol for FORESTCHECK and maybe changed to
storing each trap capture separately). Strain bulked samples through 0.2mm sieve and transfer to
70% ETOH for permanent storage in sealable glass containers enclosing sample details (site and
capture period). Sample details to be written in B or 2B pencil on durable card and placed within
storage vial.
Operator effort
It is anticipated that concentrated field sampling will be conducted for a period of 2 weeks during
each sampling season (this will achieve the hand, light trap and part pitfall trap sampling of 18
grids). Sampling is to be carried out by 2 teams of 3-4-persons. One team will place, set up and
empty light traps and Pitfall traps. For the additional two 10 day periods for the Pitfall traps,
arrangements will need to be made for assistance from Frankland District for trap opening,
closure and sample collection. This will enable flexible operation of all sampling methods
within a site. Sampling by each method need not be simultaneous on each site: different
sampling techniques can be completed on different days.
It is anticipated that sample sorting by 3-4 persons will take a further 2 weeks per sample
season. This does not include data basing.
Sampling conditions and site structure
Frequency of microhabitats within each site are recorded by walking the diagonal transect on
which the pitfall traps are set for each sampling grid. An estimate is made of % cover for: litter,
bare ground, moss and ashbeds at 8 locations along the transect. Count of stems is a count of
stems above 10 cm diam within 1m of the transect at 8 locations. A count of coarse woody debris
above 5 cm diameter is also made at 8 locations along the transect. At each site, soil and air
temperature, rain events and cloud conditions are also recorded during sampling.
Specimen processing
Specimens will be sorted and compared to a reference collection to establish their morpho-species
assignment. We anticipate using the Forestcheck reference collection for assignment of
morphospecies, however a separate WFM reference collection will be established. Specimens
>10mm only will be considered unless they have Gondwanan associations (see Forestcheck
Appendix 1). Indicator species are selected as those invertebrates which are overtly distinctive,
such that they are immediately recognizable and/or significantly notable.
Walpole Personal Equipment List for Volunteers
Here is a list of personal equipment you will require for the next 2 weeks
Essential
 Long sleeved loose working shirt(s)
 Full length trousers (loose and comfy)
 Waterproof coat
 Wide brimmed sun hat
 Socks you can tuck your trousers into
 STRONG hiking boots or steel cap boots (preferably not elastic sided; best with ankle
protection); burnt off stems are sharp and can pierce joggers easily. If you have to buy
these get one size bigger than you normally wear (you need thick socks and on hot days
your feet swell).
 A warm fleece or jumper suitable for bushwalking
 Camp shoes for resting those weary hot feet
 Towel
 Personal toiletries, medicine etc
 Personal water bottle for carrying onto site
 Pillowcase; and if you are not bringing a sleeping bag 2 sheets (the accommodation
has beds, pillows and doonas BUT you need to put you own linen over the linen
already there)
Desirable/optional
 An over the hat flynet (recommended)
 Water proof trousers
 Spare boots or good wellingtons (with thick griping soles). There is at least one plot we
may have to wade through knee deep water to get to (depends on how quickly the track
dries. (not necessaty in March)
 Camera (recommended)
 GPS or compass
 Thermos
 Signal whistle (so we can find you if you get lost on the plot)
 Gloves to protect hands when crossing slippery swamps with sword grass
 MP3 with your fave music
Leisure
 Surf board
 Swimming gear
 Fishing rod
 Chill out clothes (you will need something mossie proof for the evenings). Don't forget
this is the south coast not warm Perth and it can be cool at night.
Supplied: Hard hat, transport to and from plots; each DEC vehicle has a 2way radio and first aid
kit; and each DEC officer is trained in firstaid, plus for longer walks each DEC officer will cary a
small firstaid kit. Food; good company and stimulating conversation
There is limited 3G coverage for mobile phones. At the accomodation there is 3G coverage if
you stand on top of the 4WD (that's if we get one with a roof rack)