Good morning.
Welcome to this Natural Resources Commission webinar on fauna monitoring in New South Wales forests. Today we are focusing on yellow-bellied gliders and southern brown bandicoots, and we've got three presentations, and a panel afterwards to help answer questions on this work.
Before we start, can I acknowledge that we're all joining this webinar from Aboriginal lands, pay our respects to elders past and present, and particularly welcome anyone who is an Aboriginal or Torres Strait Islander.
On this call we have the capacity for you to ask and have your questions answered. There's a Q and A button on the top of your screen. If it isn't appearing, which does happen sometimes, then we suggest you log out and rejoin using your browser.
We'll only be answering questions about the webinar projects, and we won't be answering, or addressing wider policy issues or any specific compliance or regulatory matters. We are focusing on the monitoring work and what the messages are from that.
The questions will be moderated so that we just ensure that there are no inappropriate questions. That means there will be a small delay before your questions appear on the screen.
This webinar is being recorded and a copy will be made available on the NRC website shortly after this presentation today.
So, can I quickly introduce our presenters, first off, we've got Chris Slade, who's a senior ecologist from Hardwood Forest across New South Wales with the New South Wales Forestry Corporation, and he plays a lead role in the implementation of the Coastal IFOA fauna occupancy monitoring program. He's joined with by Leroy Gonsalves, who's a research scientist in the Forest Science unit of the Department of Primary Industries and Regional Development, and he's got a long background in wildlife ecology focusing on the responses of wildlife to forest management and disturbances in particular thinning and fire. Our third presenter is from New South Wales DCCEEW, Dr. Jane DeGabriel. She's the senior team leader, conservation science, and an ecologist with broad interest in the evolution and ecology of animal plant interactions, nutritional ecology and biodiversity. We have two panellists, after these presentations, Brad Law and Phil Gibbons, and I'll introduce them when our presentations have concluded.
So if I could hand over to Chris and Leroy, to tell us a bit about the work on yellow-bellied gliders and southern brown bandicoots.
Thank you Peter. Waiting for your slides to come up. I'll just go that one again. Can you see that, Chris? Yes. Thanks there. That's great, okay.
Thank you very much for the chance to be here today, Peter and to the NRC and also acknowledge our respect to, to Aboriginal people and the lands on which we work.
So the previous integrated forestry operations approvals, which came into being in 1999, followed a comprehensive, regional assessment and the establishment of the regional forest agreements. They considered the southern brown bandicoot as rare, and with a limited understanding of the threat posed by forestry operations. As such, a pre harvesting survey of trapping with cage traps was a requirement with a 200 hectare exclusion zone implemented around positive records. 15 of these exclusion zones were established between 1999 and 2003, and the establishment of these 15 exclusion zones provided a trigger for a review of the surveys and protective measures, and led to the establishment of additional and targeted habitat protection, which was followed by the establishment of a species management plan, in 2007.
So the species management plan, refocused the management effort into identifying and preserving Yertchuk, and scrub slash heathie dominated forest types as suitable habitat and monitoring across the landscape. And this came with the aim of understanding long-term occupancy trends and identifying environmental factors associated with the occupancy of the bandicoots in the region.
So survey program was established south of Eden on the far south coast of New South Wales at 40 sites, 20 of which were determined to be within habitat, and 20 were assigned randomly outside of this habitat. And this commenced with hair tubes during my time in Eden and up to 2009 and was pleasingly, moved to cameras, with baited lures, since 2009 with surveys conducted, in autumn and spring each year. Map there just shows you the spread of the sites across the region there south of Eden.
And like most parts of New South Wales, the program was not immune to the black summer fires, with all the monitoring sites burned, at varying intensity. This included the loss of habitat, sorry, understory habitat at 39 of the 40 sites, with the map showing the variable intensity, with black being the most intense through reds, down to greens being the lowest intensity of the fires across that landscape.
I might take over here from, from Chris. So once we've got our camera trap data and we've tagged and identified southern brown bandicoots, we look to create a metric that we can, monitor over time. And occupancy is a metric that is commonly used for landscape scale monitoring, particularly for species that occur at low densities. Um, the important thing about occupancy is that it's an approach where we can account for imperfect detection because we know that for any given method, it's not gonna be 100% we are gonna miss animals. So it's important to work out what this detectability is and then account for it when we're estimating occupancy. So things such as seasonality, for example, animals might be more active in particular seasons than others. So if we're sampling in different seasons, it's important to account for this when we are estimating occupancy. Similarly, survey methods can change, especially in long-term monitoring programs as technology improves.
Now, the process for occupancy modelling, in this monitoring program was a dynamic occupancy model, which included four steps, the first being to assess detection probability. So on a given visit, what's the probability of detecting your species given the method that you're using? Once we know what that is, and we can account for the fact that it's not 100%, we can estimate occupancy in that initial period of monitoring. So that is, what's the probability of a site being occupied by your target species. The next step in the process is to measure local colonization and extinction processes that can change from year to year. And finally, we derive a trend using that initial occupancy estimate and our colonization and extinction processes. The other thing to note is that for steps one to three in this process, we can include covariates that account for, their effects on things like detection probability.
So what were the covariates we assessed in this program? For detection probability, we included the season of sampling the camera model, because these did change over the monitoring period, the number of cameras. So sometimes you deploy two cameras, but one of the cameras would fail and also the year of survey, that was another thing that was included for detection probability. Another covariate for initial occupancy, we included a range of environmental variables including forest type elevation, TPI, which is the to topographic position, Index, annual rainfall, including one year lagged, and also some disturbance variables including, modelled habitat exclusion. So these were areas that were modelled as habitat for the southern brown bandicoot that were excluded from harvesting. We also included the extent of all areas that were unharvested within a 200 metre buffer of our sites, and similarly, the extent of harvesting of different age classes, the extent of unburnt forest, and also the extent of, burnt forests, in different age classes. Finally, we included a lidar density metric, which looked at, essentially a surrogate for, for vegetation structure at different, heights. And we also measured cat activity because we know from other work that, predator activity is, an important variable to consider for these bandicoots. And also we were getting this data on our camera trap, so it was a good it covariate to use and include for colonization and extinction probability. We included those variables that vary from one year to the next particularly those disturbance variables.
So what did we find? So when we ran our models, we found that for detection probability, season of sampling and also year of survey were the most influential variables. So detection probability varied from 0.03 to 0.4, per year, per visit, sorry. And as you can see, in 2017, that was our year with lowest detection probability at 0.03, and our highest was in 2023 with 0.4 probability of detection. For season of survey, we found that detection probability was higher in autumn compared to spring. So the key thing here is that detection probability does vary among years, and it can also vary among seasons. But with this approach, we've been able to calculate what that variability is, and we can then use that when we estimate our metric, which is occupancy at different times of the year, sorry, across different years in our monitoring program.
So what about initial occupancy? We looked at a range of variables, but there was one supported variable, and that was the extent of fire in the 15 to 30 year age class. So what we found was that sites that had burnt, in the 15 to 30, or 15 to 30 years ago had initial occupancy of 0.41, whereas sites that had burnt, sorry, sites that were unburnt 15 to 30 years ago, had a initial occupancy of 0.41, whereas sites that had been burnt in that age class had a initial occupancy that was close to zero. Now, why might that be? And we know from other work on this species that vegetation structure can be quite important. And what we found was using our lidar metric vegetation structure was, less dense in the two to four and four to six metre height classes in sites that were unburnt 15 to 30 years ago. But it was actually, more dense in sites that had burnt. So this could be a factor driving initial occupancy for local colonization probability. We found that the year of survey and also annual rainfall, including the one year lagged, annual rainfall were the most important variables influencing colonization probability. So on this plot, on the left hand side, you can see that colonization probability does vary among years. The key thing to note is this, big increase in colonization probability in the last few years of the monitoring program. And if we look at the relationship that we had with lagged annual rainfall, you can see that there was a positive relationship. So as you had more rainfall in the year proceeding sampling, you had a greater probability of a site becoming an unoccupied site becoming occupied. And just flagging again that we had our greatest colonization probability in the last few years.
For local extinction probability, we found that year of survey and annual rainfall as well as the extent of unburnt forest, which operated as an interaction, were the most supported variables. So on the left hand side, you can see that extinction probability at the local site level varied, considerably from year to year. And when we looked at the variables that might be behind this, we see that the interaction effect of, lagged annual rainfall and also whether or not the site had burnt, we found that occupied sites were less likely to become unoccupied in wetter years when conditions were good. So sites were less likely to become unoccupied, but that relationship weakened if the sites had burnt previously.
So finally, we're able to establish our trend, and what we found was that between 2009 and 2015, occupancy was between 0.4 and 0.5, but it did fluctuate among years. We saw a 53% reduction in occupancy, to 0.18 in 2016 before there was a slight upward trend observed in 2017. But the thing to note here as well is that there's high uncertainty in that year due to the low detection probability in 2017. So if you think back to that plot that I showed for detection probability, it was lowest in 2017. At 0.03 in 28 to 2020 occupancy was low. It was about 0.14, and then we saw an increase rapidly after the black summer fires and the drought breaking rain where occupancy approach close to one in 2023. And just the other thing to note here is that we didn't really see a dip following the, the black summer fires.
Now, you know, these long-term data sets are fantastic. They're giving us these snapshots in time, and we can look at how populations are tracking, but also try to assess what might be driving the trends for these populations. But there are limitations, and it's important to acknowledge these. So with this particular program, there were a small number of sites that were sampled. So we had 40 sites that were sampled, and this meant that we had low precision for occupancy estimates. The other thing that it means is that it may not be possible to sample the full range of variation in the covariates that were sampled with just 40 sites, but acknowledging those limitations, our interpretation of these results are that rainfall and fire were the major drivers of occupancy, but understory structure may also be important. As highlighted by that plot that I showed showing the, lidar results and vegetation structure, we found that black summer fires had no impact on the occupancy of this species. The extent of timber harvesting of all age classes was not associated with the occupancy trend. And similarly, the habitat exclusion zones and unharvested forests were not associated with occupancy. But we do suggest that future monitoring expands the number of sites, and as part of this process, it could be possible to establish sites with and without exclusions to provide a more rigorous test of how effective, these habitat or harvesting exclusions are. Finally, connectivity was low during the monitoring period. I didn't present this data, but it's important to note that, if connectivity does change, there's a need to, you know, implement control. But we need to continue monitoring to, to assess whether connectivity is increasing. And also, you know, are there changes in the occupancy of the southern brown bandicoot.
So just like to acknowledge, people who assisted, with this work, including, some feedback that we received from numerous people. So I'll stop there and I might hand over to Chris to, to talk about our yellow-bellied glider occupancy monitoring program.
Yeah, thanks Leroy. Great presentation. Okay. So the population of yellow-bellied gliders on the Bago platter was listed as endangered in 2008. This was based on several threats related to the power line easements, restricting movements and consideration that they were isolated with few records known in Kosciuszko National Park. At the time. Following this, a population management plan was established, for the state forests on the plateau in 2013.
So the aims of the population management plan were to establish an understanding of the long-term population trend and to assess the response of the yellow-bellied glider population to timber harvesting. The population management plan and associated monitoring program has been established in the Bago and Maragle State Forests, which are to the east of Tumbarumba on the western side of Kosciuszko National Park, at an elevation of between 900 and 1300 metres above sea level. The area receives an average of about 1,250 millimetres of rain, some of which falls as snow during winter, of course. And the forests are dominated by alpine ash, mountain gum, peppermint eucalypt species, forest types.
Yeah thanks, Chris. So I might take, take over from here and just talk people through the surveys. So there were 126 survey sites as part of this program, and these sites were sampled with core playback and one hectare spotlights between 1995 and 2021, with up to two visits to a site. From 2022 onwards passive acoustics were used to monitor the monitoring sites, but I think it's important to look at where this program has come from. So in 1995, a study by Kavanagh and Stanton surveyed 126 sites. A subset of these were resampled in 2010, and when the PMP came into place, monitoring was, part of that process. So 126 sites were sampled between 2013 and 2015, and also between 2016 and 2018. So this was, done as part of a rotating panel, assessment. So there were three panels each with 42 sites. So across the three year period, all 126 sites were sampled. So that occurred in 2013 to 2015, and also 2016 to 2018. In 2019, 40 sites were surveyed, and then we had our black summer fires. And like we saw with the southern brown bandicoot, you know, it was quite devastating to see how much of the forest had been impacted. So on the right hand side, you can see the fire extent and severity mapping, as well as the distribution of the site. So different coloured circles represent, sites in different panels, and you can see that the sites span a range of different fire severities. but the other thing to note is that 124 of the 126 sites burnt. So, you know, we had a great opportunity here to also assess the effects of the fire on Yellow-bellied gliders. So in 2020 and 2021, 51 sites were, were sampled and these were selected based on, you know, logistics. So what was accessible and safe to access immediately after the file, but also they spanned, these 51 sites span the range of forest types within the area, but also the different fire severities that affected the study sites. Now, it's important to note that we also sampled 40 of these sites also with passive acoustics to form a comparison of the two different methods. And from 2022 onwards, when passive acoustics were used, there were 52 sites sampled in that year, and, a few other sites were added in 2023. So we had 57 sites sampled in that year.
So I won't go through occupancy modelling again because we essentially use the same approach as what we did for the southern brown bandicoot. The only difference here is we had a different subset of covariates. So for detection probability, we included temperature on the night of sampling, wind, rain, the method, so whether it was spotlighting and core playback versus passive acoustics, the year of survey. For initial occupancy, again, we included a range of environmental variables, so forest type elevation, density of hollow bearing trees, annual rainfall, and then also disturbance variables, the extent of harvesting of different age classes. Also the extent of areas that were unharvested. And similarly, we did the same thing for the, the extensive burnt areas and low severity fire and high severity fire for colonization and extinction probability.
So what did we find for detection probability? We found that survey method was the most influential variable. So detection probability varied between 0.41 to 0.61, and when we had two visits to a site with spotlighting and call playback, that achieved detection probability of 0.75, whereas with passive acoustics, it was about 0.98 or 0.99, noting that we had these passive acoustic recorders out for seven nights. So across those seven nights we're able to achieve a higher detection probability. So again, like the southern brown bandicoot data, it's great that we can actually calculate what this detection probability is, but the key thing here is we know what detection probability is, and we can adjust our estimates across years when diff different survey methods are used, because then that provides a like for like comparison. Whereas if, if you ignore this then you're not really comparing apples with apples
For initial occupancy, we found that forest type was the most influential variable. So montane gum and wet peppermint gum forest groupings were more associated with occupancy. So as you had more of these forest types, you had a higher probability of a site being occupied, whereas if you had more of the western types, you had a lower probability of a site being occupied and noting the high uncertainty for, for this particular forest site, because we had a few sites that had a lot of this forest type because it was less suitable for yellow-bellied gliders. So under median conditions across all sites in 1995, occupancy was 0.17. So 17% of sites were occupied.
When we look at local colonization probability, we found that annual rainfall was the most influential variable, and that was a positive association. So unoccupied sites were more likely to become occupied in years with higher annual rainfall. And if we look at a plot of annual rainfall across the monitoring period, you can see in the last few years, like we saw with the southern brown bandicoots, that we had our greatest years of rainfall in the, in this period towards the end of the monitoring program, median colonization probability was 0.16 across all years of sampling.
What about local extinction probability? We found that the extent of high severity fire was the most influential variable, and it was a positive association such that occupied sites were more likely to become unoccupied when you had a higher extent of high severity fire within your buffer around your study site. Median extinction probability was 0.13 across the monitoring program.
So if we then look at our trend, we found that occupancy range between 0.2 to 0.3, between 95 and 2019, but it reduced by about 10% following the 2019, 20 20 fires. But we did see an increase to 0.7 by 2023. So again, like with the southern brown bandicoot data, there are limitations with these long-term data sets and it's important to identify these. So pre-fire, we had surveys with only up to two visits, to a site per survey. And this provided less precise estimates of occupancy than from 2022 onwards when we used, passive acoustics, which included up to seven nights of repeat sampling. And we can see this when we look at the trend and the tighter confidence intervals that we have for the latter part of the trend.
So the interpretation of our results are that rainfall and fire were the major drivers of occupancy. The extent of timber harvesting of all age classes were not associated with yellow-bellied glider occupancy. And this might be because harvesting is mostly focused on alpine ash, which was not associated with yellow-bellied glider occupancy, but we do know that elsewhere yellow-bellied glider occupancy is negatively associated with clear fell harvesting of alpine ash in Victoria, as well as the extent of wildfire. Passive acoustic monitoring was more effective at detecting yellow-bellied gliders in our study, and this has been demonstrated to be an effective method elsewhere in Victoria. So given this, we suggest that passive acoustic monitoring should be used for the Bago Plateau program moving forward. And finally, it's important to continue to monitor to track recovery post-fire. This is important because it allows us to assess potential cumulative impacts from other disturbances such as harvesting going forward. And I will acknowledge there were lots of people that contributed to the data sets and also provided comments and feedback along the way. So I'll stop there.
Terrific. Thanks Leroy and Chris, might hand over directly to Jane to present their work on yellow-bellied gliders. Thanks, Jane. I'm just waiting for my slides to come up. Let me try pressing that button again and while we're waiting for those, I'd also just like to acknowledge the traditional owners of all the lands we're meeting on today and pay my respects to elders past and present. And I'm joining today from Dharug and Gundungurra Country in the Blue Mountains. Hopefully you can all see my slides now.
So I'm going to be talking about a project that we did a couple of years ago looking at the impacts of the 2019/20 bush bushfires on occupancy of greater gliders and the yellow-bellied gliders across New South Wales. But today I'm just gonna focus on the results for the yellow-bellied gliders. This is work that was funded by the Commonwealth Government's Bushfire Recovery funding. And I just want to note that it was, this work was actually led by Dr. Natasha Robinson in our team, who fortunately for her is travelling in South America at the moment. And we also did this work in collaboration with Chad Beranek and Darren Southwell from the University of Newcastle. Sorry, next slide. And I just want to acknowledge the traditional owners of all of the lands that this work was undertaken on across New South Wales and pay my respects to their elders.
So, as we all know, and as Leroy has just spoken about the bushfires in the summer of 2019/20, were extreme. So the slide here gives some statistics about that. Some of the key take home ones are that about 7% of the total area of New South Wales and about 20% of the eucalypt forests in south eastern Australia were burnt. Obviously this had huge impacts for our wildlife and our boreal species, in these eucalypt forests were particularly vulnerable. And the paper written by Sarah Legge and others following the fires estimated or predicted declines of over 20, sorry, 75% for yellow-bellied gliders, which obviously is very alarming.
So we might expect that the way that a species responds to fire is going to be determined by their biological and ecological requirements, as well as the type of fire or fire regime that they experience, as well as their evolutionary exposure to fire as a threat and their behavioural responses. And consequently, we expect that species with particular traits are more likely to be affected. So, arboreal species such as the yellow-bellied glider, are likely to have really big impacts from these sorts of fires, particularly because large parts of their distribution are fire affected. So for yellow-bellied gliders, this was about 29% of their habitat. But also because of their life history and behavioural traits such as things like their limited ability to flee from the fire, low reproductive rates, and dependence on habitats that has been depleted by severe fire.
So, Leroy has just already spoken about the yellow-bellied glider, but just a quick recap of some of the traits, of some of its traits that are relevant to this study. So as we know, it's got a very broad distribution along the east coast of New South Wales, which is shown by the light grey dots on the map there. And it occurs across a very broad, elevational gradient. So from sea level up to about 1400 metres. They're dependent on hollow bearing trees in mature eucalypt forests, and they're social animals. They moved about a lot and compared to greater gliders, they have quite a broad generalist diet. So they eat things including sap foliage, flowers and arthropods, which means that they can occupy a broader habitat niche. Yellow-bellied gliders are listed, as threatened in New South Wales, and the key threats for them are associated with loss of habitat, particularly hollow bearing trees and feed trees. So we would expect that high severity fire, are going to exacerbate those threats.
So the key questions that we wanted to look at in our study were firstly, what was the probability of occupancy of yellow-bellied gliders, across their ranges in New South Wales, three years following the fires. And we also wanted to know how gliders responded to, different sub fire severity classes, across varying spatial scales. So what we predicted was that high fire severity would lead to reduced occupancy, of both yellow-bellied gliders and greater gliders, but that the yellow-bellied gliders would be more resilient due to their greater mobility and dietary breadth. And we also predicted that low to moderate fire severity wouldn't have an impact on these animals because they are able to move around across the landscape.
So in order to answer these questions, we designed a survey for yellow-bellied gliders across the range in New South Wales, and we stratified our sites according to forest types, so whether it was wet or dry sclerophyll, fire severity, so using the fire extent severity mapping, classes and elevation, so below or above 500 metres. And from this process, we ended up with a combination of about of 12 treatments, which indicated on the map there by the different coloured, triangles and squares. So in total we had surveyed 223 sites, and we ensured that those sites were spatially represented across the different treatment types across New South Wales. And we left at least one kilometre between sites For logistical reasons, we mostly conducted our surveys on National Parks estate, and also on some local Aboriginal Land Council land, on the far south coast where we worked with Aboriginal partners down there. And we tended to do our surveys from roads and tracks. We did exclude sites that had previously been intensively logged or had been burnt since 1920 so that we could minimize confounding effects. And I just want to pause here to note that unlike the long-term monitoring that Chris and Brad and Ross Goldingay and Rod Kavanagh and many others have undertaken yellow-bellied gliders, our surveys were done at a single time point. So, but when we were selecting the sites, we attempted wherever possible to include sites that had been surveyed historically, including by some of those people so that our results could feed into that longer term monitoring picture for gliders in New South Wales.
So what we did, was that we conducted our surveys over spring and summer in 2022-23, and we did three repeat surveys at each site so that we could get 80% detection confidence. We followed standard methods similar to what Leroy just talked about. We did an initial site assessment followed by 15 minutes of spotlighting along a 200 metre transect and then 10 minutes of call playback. This allowed us to develop occupancy detection models where we compared the model fit with four spatial extents. So that was at the site level, 300 metres, 500 metres and a thousand metres across three fire severity classes that was unburnt low to medium and high. And we controlled for other factors such as forest disturbance and mean temperature. This was a three step process, where we firstly developed the most parsimonious model for detection with a null model for occupancy. We then developed a base model, which was the most parsimonious model for occupancy and detection, excluding the fire hypothesis variables. And then we compared the base model and with each fire covariate to get the best overall model.
So the results of our occupancy analysis for yellow-bellied gliders is shown here on the map with the red dots where we found yellow-bellied gliders. And essentially what we found was that the three years following the fires, yellow-bellied gliders had a relatively low occupancy, with a occupancy score of about 0.07, which I was just pleased to hear was quite consistent with what Leroy found in Bago. We only detected them 54 times across 610 surveys, and they were only present at 31 out of the 223 sites that we surveyed. One thing that we did find, though, was that even though occupancy was low, we did find, an increasing probability of detecting yellow-bellied gliders with increasing temperature across all sites.
So now looking at our models, the top performing, base model for the yellow-bellied glider included one detection covariate, which was negative, sorry, positive, and that was temperature, and two occupancy covariates. So that was the ecosystem disturbance index layer, and annual temperature variation. And they were both negative. So the figure on the right here shows the results at the site scale. And we found that the local FESM score was the most predictive fire covariate for yellow-bellied glider occupancy at this scale. But when we considered across all sites, we found that yellow-bellied gliders had, a 2.2 times higher likelihood of occupancy in unburnt sites, and sites with low, to moderate severity burns had a 2.5 times higher likelihood of occupancy compared to sites with high severity burns. So those high severity fires were the most important predictor there, which again, is consistent with what Leroy just presented to us. And we found that low to moderate severity fire buffers were less predictive at all spatial scales compared to our base model that had no fire covariate.
So kind of just summing up and what this means, our results show that the yellow-bellied gliders five severity was the most important at the site scale, and populations declined at severely burnt sites. And this suggests that species traits are important in shaping the responses of the yellow-bellied gliders to fire at a landscape scale because due to their high mobility and their relatively large home range sizes, three years following the fire they were actually able to move around and seek out unburnt resources within their nightly foraging patterns. Our analysis found spatial heterogeneity in fire impacts across the landscape. So there was a matrix of unburnt and burnt areas, which increased the likelihood of yellow-bellied gliders being able to access those unburnt resources within their nightly foraging and actually be able to find the resources that they need.
So just to sum up what this all means, our results indicated that yellow-bellied gliders were particularly vulnerable to high severity fire, but resilient to low to moderate fire, which is what we would expect based on their evolutionary history. I didn't go into the results of the yellow-bellied gliders, sorry, greater gliders today. But just to note that our results did find that the yellow-bellied gliders were more resilient than greater gliders to severe fire at the landscape scale. We found that yellow-bellied gliders preferred unburnt sites, but severely burnt forests wasn't the barrier to movement including immediately following the fires. And our results indicate that severe fire is expected to impact on habitat quality in the short term, but with increasing time since fire forest recovery is likely to help facilitate the movement of gliders through severely burnt areas.
So these results can help us to develop management recommendations for yellow-bellied gliders. In particular, it's clear that we need to minimize the risk of landscape scale high severity fire, and we can look at doing this, through things such as, ensuring low to moderate fire, through things like cultural burning or appropriate control fire to minimize the risk of future intense wildfires. And it's really clear that we need to maintain essential habitat for yellow-bellied gliders, particularly ensuring the conservation of large mature hollow bearing trees and feed trees in the landscape. As I mentioned before, our results were done at a single point in time following the fires, but these results can be built into these other studies as well to help build up and inform our monitoring design for future yellow-bellied glider monitoring surveys. And I just wanted to finish by giving a quick plug to a publication on this work that will be coming out soon in wildlife research. And I just want to thank all of the people, many, many people that were involved in helping us conduct these studies, including many of our National Park colleagues and local Aboriginal rangers, and thank you all, and I'll stop presenting now.
Thanks very much, Jane. I'm sure that has prompted a number of questions. I've noticed we have one question already on Q and A, so if you've got questions, please start loading them into the Q and A session. At the moment though, we've moved to our panel, we have two panellists, Brad Law, who's a principal research scientist for the Forest Science Unit in DPIRD and Phil Gibbons who is a professor at the Fenner School at ANU, both who've been working a long time on wildlife and conservation issues in forests. Phil is a independent expert advisor to the New South Wales Forest Monitoring Steering Committee. And with that brief introduction, I might turn perhaps to Phil first as to whether you've got any first reactions or questions. And again, to the audience, please start populating the Q and A panel. Thank you.
Thanks Peter. Look first on the southern brown bandicoot, thanks to the speakers, that was really clear and fantastic. From my perspective anyway, look so the southern brown bandicoot, I was on a technical working group on that species, and as we watched that decline in abundance, we got very concerned. And, you know, there's been sort of examples of where people just watch species decline to extinction. So we really thought about, okay, we need to do something here. This was around about 2020 and we started to think about how to design predator control, because there was no evidence that actually the species was declining any worse in harvested forest versus unharvested forest. So we really shifted to predator control as an option. But the rains came and that species bounced back incredibly. It was amazing how it bounced back. And I take my students down to Eden and we get them on cameras quite regularly, not just the southern brown bandicoot, but some of the others as well in that critical weight range. And we found a similar thing after the Orroral Valley fire and the Namadgi National Park, a similar response by, long-nosed bandicoot. So that was amazing. On the yellow-bellied glider, it does concern me. You know, there's a common thread in both of those studies, and that is high severity fire is an issue for this species. And you know, we know that the period between a fire interval is decreasing in New South Wales forest with climate change. And while it hasn't been found categorically, you would think that with the increasing fire, severity of fire weather, that we're going to get larger areas of more severe fire. So this is a real concern. So how do we make our forests more resilient to more intensive fires, is something we really, really need to think about. And we know also in some of the work that I've done, with colleagues in that high severity fire is an issue with a decline of mature trees. And as long as you've got that mosaic of lower severity fire and high severity fire, you don't get a net loss of mature trees at the landscape scale. But within those areas of a high severity fire, we have an issue. So that is something that really we need to think about in future. So I'll leave it there. Thanks Peter.
Thanks Phil. Brad, I turn to you before we start looking at the q and As. Yeah, just a few quick general comments. I suppose I just wanted to really highlight the importance of these sort of long-term programs. There's not a whole lot of them out there, and I think they're just super important to look at the changing status of species over time. And you know, it can take 10 years before we've got a clear picture on what those trends are and what the drivers are. So it does take time for those trends to come through. And yes, it's very important going forward to be used to look for the effects of management, changing fire regimes and of course climate change. Over the top of that, I think they also present a few other opportunities, especially with some of the methods now being used that other species can be analysed, from these programs. So data is from the camera traps has also captured long-nosed potoroo, so there's the possibility of looking at their trends over time and using acoustics for yellow-bellied gliders, that's also capturing information on forest owls. So that's very, you know, really useful. And then just lastly, we probably wouldn't have predicted these responses of these species if we hadn't actually have gone out with targeted monitoring of species. If we relied on condition metrics, perhaps recorded via remote sensing, we just wouldn't have had the same appreciation of how these species have responded over time. So I think that's something important to think about.
Great. Thanks, thanks Brad. We've got a few key questions coming in, which I'll turn to in just a second. One quick question for all of you though is, we use the term unburnt, and I know Leroy defined that effectively, I think as 15 to 30 year period of fire free, but there's probably some degree of variation in length of time since fire, and whether we just use the term unburnt as a comparator might need a little bit more finessing, I suspect.
Let me, our first question that's on the top of the pile, has come in from Clayton Sparke, and if I can just read that to you, if you can't see it, it says, Leroy and Chris touched on the effect of changing technology in long-term monitoring programs. Can you share any details on the type and influencing influence of new technologies that were implemented over the course of the programs or insights into emerging technologies? You did touch on that Leroy and Chris.
Yeah that's right. That's a good question, Clayton. So I guess like thinking to like the yellow-bellied glider study, for example, technology has shifted where, you know, we now have access to these passive acoustic devices, so they're, you know, I guess more readily available. But we also have tools to help us process the data that, these sensors collect. So they collect lots of information. So obviously when you're out there spotlighting and doing call playback, it's for a limited point in time. So these devices are out all night, multiple nights. So as we saw with those results detection probability, was higher with the longer deployment, with these passive acoustics and, you know, the technologies are only going to get better and better. You know, people are using drones to monitor some of these species and collecting, you know, fine scale detail, you know, may maybe even density. So yeah, I can’t, where we'll get to in 10 years, I can't really imagine. But yeah, I don’t know if anyone else on the, panel has any thoughts on that.
Yeah, I might also just jump onto that. Leroy, you touched on the drones and that's something that we are looking at the moment. So we've got a new grant from the Environmental Trust to use drones to look at that really intensive site scale to develop really calibrated methods for drones, for glider surveys, and compare those to the spotlighting methods so that we can actually understand where our limitations are and develop really targeted methods for surveying, gliders, but also using LIDAR attached to the drone survey bat habitat so we can match up some of those characteristics and get that more detailed understanding. And just, just another quick comment as well, for the southern brown bandicoot, that was obviously camera trapping data and over the past, you know, 10, 20 years, the models for camera traps have improved greatly. So when the program started, they were fairly primitive cameras, but now they're, you know, they're much more effective. But as Leroy said, that's taken into account, when modelling detection probability, but also there's AI development in that space as well. So at the moment, people mostly go through manually to identify images, but that's progressing and not too long in the future that there'll be some automated process to help out. And I'm sure we'd all agree streets ahead of hair tubes. Absolutely.
Teresa has said, or asked for both the yellow-bellied and greater glider surveys were site-based key habitat features, also surveyed like large hollow bearing trees and sap feed trees. Very concerned about the attrition of these features. Jane? I guess I might answer that, hello Teresa Unfortunately no. So with the resources we had available, it was just already so intensive to be able just to do the spotlighting surveys at that number of sites. And we weren't able to do those sort of daytime surveys as well. But I would love to be able to do that and I think that's the sort of thing that we would like to think about in future as we continue our research program. But yeah, really good point.
We've got a great question from Ruben who's a farm forester on the south coast, although he's asking primarily about PNF private native forest arrangements. But he's got a great question here about high density regrowth following past disturbance under previous owners. And he was keen to see the health of his forest develop to the extent it can become home to diverse species and he does have a specific point about exclusion zones for SBB and greater gliders and their impact on viability. Does anyone want to tackle that? That might be a Chris one, maybe.
Sure, Peter. Yeah, so there's a range of, protection measures that are employed in the, at least the public forests. And they relate to den and hollow tree protection as well as we all understand the significance of those, Sap feed trees. They are repeatedly used. So those particular features are are very well protected and identified across the public forest at least, and I'm sure the private native forestry work as well. So they're recognizing those particularly important habitat features are identified. And you know, I guess well recognized that they are pretty significant in the whole ecology of these species. And particularly I think as Leroy pointed out too, the habitat for the bandicoot, was understood to be reasonably dense sort of understory in various places over the time. But turning out maybe it's not quite the case, with the results now, but those aspects are, really important to be managing across the landscape for all these species as well.
Great. Thanks Chris. Ruben asked a couple of other questions which will probably get answered I think out outside of this particular forum, but there are some great questions there. I noticed someone's got their hand up, but can you please put to your question into the Q and A so that if we can't get to it, we can answer it after the webinar's finished? Kevin has asked Jane, I'm curious whether the yellow-bellied gliders are put off by charcoal on severely burnt trees. Do they need patches of unburnt bark to set up feed sites on a tree? Good question, Kevin I'm not actually sure about the answer to that. I don't, I'm not an expert on yellow-bellied guider feeding. I would imagine that it's, it depends on the severity of the burn. So I would say that if a tree was severely burnt, then they probably wouldn't be able to get the sap out of the charcoal. But I don't know the exact answer. I feel like Ross Goldingay would probably be our best person to answer that question unless somebody, Brad or Chris have got any thoughts on that. My only comment would be that I guess they're often probably not exclusively, but often feeding on smooth bark trees, which obviously are going to have some charring after the fires, but a lot of the smooth bark trees were shedding their bark fairly quickly after the fires within one to two years. So probably that charring effect may not, if the tree's still alive, may not persist for too long, at least on the smooth box.
Thanks. Ian Parsons asked, what camera trap settings would you recommend for best results with yellow-bellied glider monitoring? I might try here. So I was going to say that, you know, we haven't really done, you know, camera trapping for yellow-bellied gliders as part of, this monitoring program. But I was just going to flag that, acoustics is something that we have used and a big part of that technology being able to process these large data sets. And we have developed recognisers particularly for yellow-bellied gliders. So if people are interested in, starting their own programs using acoustics, we have that available on our website so it's freely available for people to download and use. And you know, there are some caveats, but that's all kind of presented on the website. So yeah, just thought I'd flag that. But yeah, if anyone else has thoughts about camera trapping, for yellow-bellied gliders, Jane, Brad, Chris? No, okay.
We're getting to questions that were actually asked first up. So Frank Lambert asked what, he's just possibly a naive question, but could animals need to range further to find resources after fires and therefore become more likely to encounter and be recorded on a recorder than before the fires? So any thoughts on that? Looks like it might be a good question. It's certainly something to be teased out, sorry Leroy, it's certainly something to be teased out, Frank. I think it is a good question and not naive at all. I think there's a lot to be learned in that space, that is a real potential opportunity there. Definitely. Yeah, I was just going to add to that, if that is the case, what might be behind that, might not be clear. But, you know, going back to the modelling, the key thing is knowing that if there is a variation before and after fire in detectability, if animals are moving further then, we are calculating that and accounting for that when we are estimating occupancy. So, you know, if we don't know what might be driving that, maybe it's not critical for, you know, monitoring that trend in occupancy with that being the metric.
Okay. Fantastic. Thank you. I've got two more early questions before we go back up to the top. So Mary Rose Daoud, they asked how does acoustic surveys or how do acoustic surveys differentiate between individual gliders species or individual gliders? I'm sure that she means individuals as opposed to species. Yep. Yeah, so that's a great question. And the method at the moment doesn't differentiate between individuals. So it will detect a call of a yellow-bellied glider and these metrics that we report on are occupancy at the species level. So it's not telling you anything about individuals. But you know, like we said, in 10 years, 10 years time technology could improve where we could get to that point where we're identifying individuals. So we'll have to watch that space Possible job for artificial intelligence tools.
Ben Hope has asked for the southern brown bandicoot. Is it possible that the exclusion zones function as refugia and allow for recovery outside of the exclusion zones when conditions allow? That's his first question, and then he is asking about dogs. Let's take the question about exclusion zones As refugia. I guess I'm happy to jump in there, Peter. I think yes, that would be definitely the case, to act as refugia. Definitely, yes. and I know, dogs are monitored as well as part of the cats and foxes in that part of the world too. So there are, known wild dog populations slash dingo, in that part of the world that obviously the species falls within the critical weight range as Phil described before. So, they're able to be preyed upon by those groups of species. So that refugee is also important for that aspect as well. Yep. Just one, one more comment on that, you know, cameras are obviously monitoring dogs, foxes and cats at the same time as the bandicoots and Leroy can correct me, but I think there were very few detections of dogs and foxes, but, and cat's activity was relatively low, but, even lower for foxes and I think dogs.
Okay. Thank you. I've got two more, three more questions here. Steven Dobbyns, three minutes ago asked, more a comment than a question. Having seen the landscape wide severity of the bushfires across the Kosciuszko region. I doubt we'll see gliders for many decades due to the lack of hollow bearing trees. Some more detailed question here, just about land management, Chris, or, well, Phil, Yeah. Can I just make a comment on that, Peter? I mean, we've done work, You know, the thing is that fire is an agent of hollow destruction and creation. So it's not, it's not all negative with, fire and hollows. So, high severity fire can be an issue, but at the landscape, but low severity fire actually increases the number of hollows, a number of studies including our own have found. So as long as you've got that mosaic of low and high severity fire, then you're not necessarily getting a net loss of hollows a number of hollows in the landscape. And that's, you know, that's born out in the relative sensitivity of the yellow-bellied glider that ranges across a wide area relative to the greater glider, which, Jane mentioned just briefly as well and their relative sensitivity. So yeah, so it's the fire, you know, these things have evolved with fire. It's the changing fire regime that might be an issue if we start to get a greater percentage of the landscape more severely burnt.
Thanks, thanks. Phil, Bethany Dunn has asked as passive acoustic monitoring progresses, it would be great to liaise across jurisdictions in calling call recognizes and methodologies, so we don't all reinvent the wheel, yes. cause all of this work is based in New South Wales, so collaboration across borders, between research groups. I see Jane nodding ahead. Yeah, couldn't agree more, Bethany. In fact, our team is currently working with the researchers in Victoria from the Arthur Rylah Institute. And also Paul Roe and his group from James Cook University in Queensland around acoustic monitoring at the moment. Something we are really interested in, particularly in relation to the Saving Our Species Program. But always happy to talk to our, cross jurisdictional counterparts to, you know, compare what we're doing and talk about these things. So if you want to get in touch with me, send me an email and my team can have a chat to you about that.
Terrific. I think our last question here is, Patrick Tegart, from Wingecarribee Shire who are doing drone surveys, in the council managed Bushland Reserve, which has greater gliders and yellow-bellied gliders, both detected by drones. He's offering to reach out as well. Looks like you've got a, another collaborator there. Excellent. Look forward to hearing from you, Patrick. I'll send you an email after this and we can get in touch. Thank you.
And, Shu Lee has asked the great question at the end, which is yes, the webinar has been recorded, and will be available shortly on the NRC website. And there will be a closing slide, which gives you, a link for that to find out more information on the program. The two studies that Leroy and Chris referred to, the reports from those studies are actually on the, NRC website under latest news. Jane, I think you referred to your work being published shortly, which is great. And we have a QR code here for feedback on the webinar. We are particularly interested in your feedback, so I encourage you to follow that through. If not the QR code, you will get an email with a link as well. We're keen to plan and deliver an informative and topical webinar series, and so your feedback is really valuable. If I could thank now all our presenters, Chris, Leroy, Jane. And Phil and Brad for joining us on the questions. Look forward to seeing you all next time. Thanks for joining and your interest.