Developing user-friendly techniques to monitor arboreal geckos
New methods are currently being developed for monitoring the highly cryptic forest gecko in biodiversity sanctuaries. Image – Trent Bell.
Lizards comprise around half of New Zealand’s native terrestrial vertebrate species. Of these, 68% are considered threatened or at risk, predominantly due to predation by introduced mammals and habitat loss. All invasive predatory mammals prey on lizards. Two lizard species have become extinct, and more have suffered extensive range contractions and fragmentations. Others remain widespread but persist at lowered densities on the mainland.
Biodiversity sanctuaries may be able to help reverse this decline. However, how arboreal and nocturnal geckos respond to pest control is poorly understood, largely because no user-friendly monitoring method currently exists for these geckos. Arboreal geckos are small, secretive, well-camouflaged inhabitants of complex habitats and are difficult to monitor.
To address these issues, a research programme has been set up:
- develop robust monitoring methods which can be used by sanctuary staffand volunteers for nocturnal arboreal geckos, and
- study the responses of these geckos to pest management in sanctuaries.
In winter 2010, 200 closed-cell foam retreats (CFRs) were placed on tree trunks inside a non-fenced sanctuary in which rodents and stoats are baited and trapped. Possums and weasels are absent and geckos persist in reasonable abundances. A further 200 CFRs were placed in reference areas (no pest control) surrounding the sanctuary. All CFRs were placed in secondary kānuka forest. In March 2011, the CFR's were checked every second day over 12 days and the following data recorded: the number of geckos per CFR, their sex and snout-vent length (SVL), and weather covariates (temperature, relative humidity, cloud cover, and wind strength).
To analyse the data collected occupancy modeling of gecko presence/absence in CFR's was trialled as an indicator of abundance. Occupancy modeling is a relatively new technique that is well suited for monitoring rare or cryptic species. Because it infers detection probability from the detection history of each CFR, individual geckos need not be identified, and the sampling effort required is usually predictable and manageable.
Single-season models were run in the computer program PRESENCE 4.0 to estimate occupancy of sanctuary and reference CFRs. While estimated occupancy of reference CFRs was about half that of sanctuary CFRs, the difference was not statistically significant (Fig. 1). However, simulations based on the estimated occupancies and averaged detection probabilities indicated that the difference would have been significant had eight checks of reference and six checks of sanctuary CFRs been performed. Detection probability of geckos differed each day, between 4.2% and 45.9% for sanctuary CFRs and 4.5% to 40.6% for reference CFRs, highlighting the difficulty of detecting forest geckos.
For every female gecko found under CFRs, 2.2 males were found but this male:female ratio did not differ significantly between sanctuary and reference sites. Nor was there a significant difference in the sizes of geckos in sanctuary and reference CFRs, although noticeably more geckos with SVLs between 81 and 95 mm were found under sanctuary CFRs (Fig. 2). The sanctuary may have more large geckos due to improved survival rates following pest management.
Data collection for indices of animal occupancy or abundance is less resource-intensive than for estimation methods, but indices can be unreliable because they do not incorporate detection probability. Five occupancy and abundance indices were calculated from the gecko data for comparison against occupancy estimates (Fig. 3). Encouragingly, all indices performed similarly to the occupancy estimates (Fig. 3), suggesting that indices can provide a reasonable indication of gecko abundance.
The weather conditions affected CFR use by geckos. They were used by fewer geckos when it was warmer and more humid, but by more geckos when it was windier and cloudier. Therefore, it is best to check CFRs when it is cool, dry, overcast and breezy.
Occupancy estimation of CFR use proved to be an appropriate method for monitoring forest geckos. Using these models, it was stimated that twice as many geckos occupied sanctuary compared to non-managed CFRs. Detection probability was very similar between reference and sanctuary CFRs suggesting that geckos benefi t from intensive pest control. When resources are limited, index estimates could also be used but validation against occupancy or mark-recapture estimates is advisable. The challenge now is to further develop the model using habitat covariates and apply this method to other study locations and species.
This research was funded by the Ministry for Science and Innovation (Programme COX0903J) and Auckland Council.
Trent Bell
Contact Trent: trent@ecogecko.co.nz
Sarah Herbert (EcoGecko)
John Innes (Landcare Research)
Matt Baber, Ali Thompson & Su Sinclair (Auckland Council)
Fig. 1 Estimated occupancy ( ± 95% CI) of forest geckos under CFRs.
Fig. 2 Snout-vent lengths of forest geckos found underneath CFRs.
Fig. 3 Comparison of index methods for forest geckos underneath CFRs with estimated occupancy. 1st day abundance = number of geckos caught during the first check; Mt +1 = number of individual geckos caught over 6 checks; occupancy index (1 day, 3 day and 6 day) = cumulative percentage of CFRs occupied without estimating detection probabilities.
