I would not be 'confident' in making a determination that the condition of a site had improved based on aerial imagery only. Aerial imagery is best used to indicate changes in vegetation cover and landscape processes such as flooding & fire. You could assume that the more extensive shrub layer and ground cover in the 2010 image indicates a healthier vegetation community. However, ground-truthing the site may in fact reveal the assumption as false: the shrub layer may have been invaded by Cassinia; ground cover is dominated by weed species that are inhibiting recruitment of native species; and the dominant canopy species are even-aged and trending towards senescence.
Biodiversity monitoring must assess a range of site-based attributes that are relatively easy to measure and ecologically robust. To focus on the measure of condition for a vegetation community, you can’t isolate a sample site from the range of spatial and temporal influences at the wider landscape scale - its context and connectivity and relationship with threatening processes.
If the goal of this exercise is to make a meaningful comparison between the same site over time (for the purpose of biodiversity conservation) then I believe you should be sampling in the same season. For instance, you can experience a markedly higher native grass and forb species richness in spring/summer in a period of relatively higher rainfall. It would not be valid to compare this sample with one taken in a period of drought or in a season of dormancy.
The next question to answer: is there a correlation in condition improvement over the 50 years with management actions e.g. grazing exclusion and removal of exotic species; or is condition determined by natural cyclic processes?
Matt Colloff - 12 years ago
If the condition of vegetation fluctuates according to interdecadal wet and dry phases, then the variation in condition is a response to a natural climate cycle of the region. It follows the vegetation that exists in that landscape is adapated to some degree to long-term oscillations in wetting and drying. Does it then make sense to say the condition associated with with 'wet' is 'good' or that associated with 'dry' is 'bad'? Is it not the drought-stressing of trees and the associated 'dieback' of large limbs and branches that creates tree hollows and coarse woody debris that provide habitat and resources for other organisms? Our concept of what is good and bad in relation to vegetation condition starts to become a bit blurred when we consider the broader implications for other parts of the ecosystem. And there in lies a possible modified approach; the consideration of condition in relation to the consequenses for other organisms. Does this all mean that it is a waste of time monitoring vegetation condition? Of course not. We want to know whether drought cycles and the vegetation response is being accentuated by climate change and water diversions. We want to know where the points of irreversible ecosystem change may lie and whether there are effective interventions we can make. But first we have to redefine what is 'better'.
Peter - 12 years ago
A good discussion prompted by your post Ian, one of many in a very informative and thought provoking blog.
I would endorse the approach proposed by Chris above wherein two points in time alone will not be sufficiently informative to determine the trend. As has been evident with the breaking of the last drought in eastern Australia, recovery in vegetation condition affects different components of the community and its structure over different periods of time. What is needed therefore to determine the trend is an understanding of the context of the site and the position in the cycle of recovery, senscence or other types of disturbance.
Similiarly a more technical analysis of the aerial imagery, particularly if a greater number of comparison sites were considered, would give an understanding of comparative vegetative cover and allow for an understanding of the changes over a larger area in order to inform conclusions about the specific site of interest. Further to this, there is also likely to be a bias in the interpretation of the early photograph given that it is in black and white and somewhat distorted. This would affect perception of crown dimensions and other features. Processing the images in a manner that makes comparison more equitable may also assist.
If two points in time are a limitation of the available information, there needs to be an understanding of the preceding conditions and to have equally comparable images from the outset.
Cheers,
Peter
Tim Barlow - 12 years ago
Hey, there’s no option for ‘all of the above’!!
But here's my thoughts (and I agree with much of the other commentary above).
If the question is “has the condition changed?”, then I don’t think it matters whether the antecedent seasons are similar or not - we’re not asking *why* changes have occurred, just “is condition better or worse?”. A Govt funding agency may like this approach, since if there has been an improvement it can be claimed that it is due to Govt funding. If there hasn’t, it is most likely due to inefficient use of funds by regional delivery groups (ok, so I wear a black hat!).
However, I imagine the some of us would want some idea as to why, not just if, condition has changed. In that case, we would want to know what’s happening with a number of influential variables, especially if we can’t 'manage' them. In addition to seasonal conditions, we may also want to know what’s been happening with respect to fire, grazing, surrounding land-use and so on. Analysing effects on condition would be easier if there is less variation in these parameters (or ‘controlled’ – e.g., the site experienced 5 years of drought immediately prior to both assessments). But as these variables are difficult to control, and resources to measure are presumably limited, any evidence from the literature or other sources as to how they might influence condition could be useful in describing changes.
I’m a bit worried about having to make a decision on whether the site has improved or declined on the basis of just two assessments – ideally, I would like further detail on what has been happening in the intervening period. It may be that short term influences (e.g., severe drought) are masking longer-term trends, but I’ll take this as limitation which we have to deal with in our decision making (again, calling on the available evidence).
So:
Option 1 is desirable, but drought (or whatever) is only one of many variables which could be controlled in the sampling sense, so concentrating on it alone would be a mistake;
Option 2 is valid, providing the influence of climate (and other variables) is factored in to the assessment by calling up relevant evidence (if indeed it exists);
Option 3 is also valid, although with the qualification that neither are the most appropriate. But both are useful and have limitations that are manageable;
Option 4 looks attractive; I don’t know what the best approach is exactly, but I’ll bet you’ll never get the funds to implement it!
On this basis, I have gone with #3: either approach could be useful, used skilfully.
Gregg Muller - 12 years ago
An interesting question Ian. I concur with Cathy's comments about unpacking 'vegetation condition'. Poor veg condition in one location might provide opportunities for unusual organisms to utilise that area - or nutrients might be transferred off site, benefiting another location. There might be spatial or temporal coupling going on at scales that we aren't picking up on. And at the risk of getting all philosophical, 'condition' seems to necessitate conceptions of good and bad - but good for who? good for what? good for when?
I enjoy the before and after photos. But I have to admit that I doubted whether I could look at two aerial photos of ecosystems I was unfamiliar with and judge which was healthier. I was happy to accept your judgment, Ian, and try to learn from it.
I voted that both approaches could be appropriate. If I had a choice, I'd certainly prefer to compare wet years with wet years, year after fire with year after fire, etc. But as Chris and Cathy point out, there are many variables changing, and that difficult challenge is likely to accelerate. We need to develop techniques that allow us to evaluate change over time despite the "noise" of changing herbivores, invasives, climate, diseases, predators, successional pathways, etc. etc.
In my own conservation monitoring I've found the Floristic Quality Assessment (Gerould Wilhelm, John Taft, Doug Ladd, etc.) to be especially helpful. Previously standard measures of diversity, "evenness," etc. seemed much less useful in measuring changes in "health" from a biodiversity conservation perspective.
The basic principle of the FQA is that our conservation goal should be a high or rising diversity of conservative species. It is relatively less "distracted" by changes in rainfall, season, and other short-term quirks.
An interesting discussion. One way to address the issue is simply to recognize that two points in time are not sufficient for measuring change in ecosystems. Year to year fluctuations in management, climate, and other factors can alter some attributes of those ecosystems pretty dramatically, but measuring them at a single point in time only shows where the ecosystem is within its range of variability. The bigger question is whether that range of variability has changed - has it shrunk, expanded, or moved in one direction or another? In other words, you really need to be looking at a running average over time and looking for long-term trends.
As an example, herbaceous vegetation structure can be very different based on herbivory rates. Those herbivory rates might be driven by stocking rates of livestock or by population cycles and by rainfall - which alters the amount of total forage available. Just because the structure is short one year doesn't mean it won't be tall the next. If vegetation structure is the most important variable to assess, it's obviously not fair to assess the capacity of the ecosystem to create vegetation structure by looking at it only in a year of drought stress and/or high stocking rates of herbivores.
Excellent discussion.
Chris
Cathy - 12 years ago
I think not just weather but consider any variable that impacts on ‘condition’....and taking a step back and the notion of ‘vegetation condition’ may need unpacking. Condition for what? may be a question. Think of change, evolution (long and slow) or natural disasters (episodic and rapid) – prevailing conditions advantage some species or ecosystems, whilst wiping others out. If there is increasing variability, as we’re expecting with the impacts of climate change then we may need to revisit what we think of as ‘vegetation condition’ and how we grade it. What would you think about differences like natural seasonal variability? Could we think differently, temporal condition maybe? Which then moves my mind towards resilience theory and back loops and I’ll stop here.
My 2 cents is up.
Thanks for your postings Ian, keeps my brain cogs a-whirring.
Best
Cathy
I would not be 'confident' in making a determination that the condition of a site had improved based on aerial imagery only. Aerial imagery is best used to indicate changes in vegetation cover and landscape processes such as flooding & fire. You could assume that the more extensive shrub layer and ground cover in the 2010 image indicates a healthier vegetation community. However, ground-truthing the site may in fact reveal the assumption as false: the shrub layer may have been invaded by Cassinia; ground cover is dominated by weed species that are inhibiting recruitment of native species; and the dominant canopy species are even-aged and trending towards senescence.
Biodiversity monitoring must assess a range of site-based attributes that are relatively easy to measure and ecologically robust. To focus on the measure of condition for a vegetation community, you can’t isolate a sample site from the range of spatial and temporal influences at the wider landscape scale - its context and connectivity and relationship with threatening processes.
If the goal of this exercise is to make a meaningful comparison between the same site over time (for the purpose of biodiversity conservation) then I believe you should be sampling in the same season. For instance, you can experience a markedly higher native grass and forb species richness in spring/summer in a period of relatively higher rainfall. It would not be valid to compare this sample with one taken in a period of drought or in a season of dormancy.
The next question to answer: is there a correlation in condition improvement over the 50 years with management actions e.g. grazing exclusion and removal of exotic species; or is condition determined by natural cyclic processes?
If the condition of vegetation fluctuates according to interdecadal wet and dry phases, then the variation in condition is a response to a natural climate cycle of the region. It follows the vegetation that exists in that landscape is adapated to some degree to long-term oscillations in wetting and drying. Does it then make sense to say the condition associated with with 'wet' is 'good' or that associated with 'dry' is 'bad'? Is it not the drought-stressing of trees and the associated 'dieback' of large limbs and branches that creates tree hollows and coarse woody debris that provide habitat and resources for other organisms? Our concept of what is good and bad in relation to vegetation condition starts to become a bit blurred when we consider the broader implications for other parts of the ecosystem. And there in lies a possible modified approach; the consideration of condition in relation to the consequenses for other organisms. Does this all mean that it is a waste of time monitoring vegetation condition? Of course not. We want to know whether drought cycles and the vegetation response is being accentuated by climate change and water diversions. We want to know where the points of irreversible ecosystem change may lie and whether there are effective interventions we can make. But first we have to redefine what is 'better'.
A good discussion prompted by your post Ian, one of many in a very informative and thought provoking blog.
I would endorse the approach proposed by Chris above wherein two points in time alone will not be sufficiently informative to determine the trend. As has been evident with the breaking of the last drought in eastern Australia, recovery in vegetation condition affects different components of the community and its structure over different periods of time. What is needed therefore to determine the trend is an understanding of the context of the site and the position in the cycle of recovery, senscence or other types of disturbance.
Similiarly a more technical analysis of the aerial imagery, particularly if a greater number of comparison sites were considered, would give an understanding of comparative vegetative cover and allow for an understanding of the changes over a larger area in order to inform conclusions about the specific site of interest. Further to this, there is also likely to be a bias in the interpretation of the early photograph given that it is in black and white and somewhat distorted. This would affect perception of crown dimensions and other features. Processing the images in a manner that makes comparison more equitable may also assist.
If two points in time are a limitation of the available information, there needs to be an understanding of the preceding conditions and to have equally comparable images from the outset.
Cheers,
Peter
Hey, there’s no option for ‘all of the above’!!
But here's my thoughts (and I agree with much of the other commentary above).
If the question is “has the condition changed?”, then I don’t think it matters whether the antecedent seasons are similar or not - we’re not asking *why* changes have occurred, just “is condition better or worse?”. A Govt funding agency may like this approach, since if there has been an improvement it can be claimed that it is due to Govt funding. If there hasn’t, it is most likely due to inefficient use of funds by regional delivery groups (ok, so I wear a black hat!).
However, I imagine the some of us would want some idea as to why, not just if, condition has changed. In that case, we would want to know what’s happening with a number of influential variables, especially if we can’t 'manage' them. In addition to seasonal conditions, we may also want to know what’s been happening with respect to fire, grazing, surrounding land-use and so on. Analysing effects on condition would be easier if there is less variation in these parameters (or ‘controlled’ – e.g., the site experienced 5 years of drought immediately prior to both assessments). But as these variables are difficult to control, and resources to measure are presumably limited, any evidence from the literature or other sources as to how they might influence condition could be useful in describing changes.
I’m a bit worried about having to make a decision on whether the site has improved or declined on the basis of just two assessments – ideally, I would like further detail on what has been happening in the intervening period. It may be that short term influences (e.g., severe drought) are masking longer-term trends, but I’ll take this as limitation which we have to deal with in our decision making (again, calling on the available evidence).
So:
Option 1 is desirable, but drought (or whatever) is only one of many variables which could be controlled in the sampling sense, so concentrating on it alone would be a mistake;
Option 2 is valid, providing the influence of climate (and other variables) is factored in to the assessment by calling up relevant evidence (if indeed it exists);
Option 3 is also valid, although with the qualification that neither are the most appropriate. But both are useful and have limitations that are manageable;
Option 4 looks attractive; I don’t know what the best approach is exactly, but I’ll bet you’ll never get the funds to implement it!
On this basis, I have gone with #3: either approach could be useful, used skilfully.
An interesting question Ian. I concur with Cathy's comments about unpacking 'vegetation condition'. Poor veg condition in one location might provide opportunities for unusual organisms to utilise that area - or nutrients might be transferred off site, benefiting another location. There might be spatial or temporal coupling going on at scales that we aren't picking up on. And at the risk of getting all philosophical, 'condition' seems to necessitate conceptions of good and bad - but good for who? good for what? good for when?
I enjoy the before and after photos. But I have to admit that I doubted whether I could look at two aerial photos of ecosystems I was unfamiliar with and judge which was healthier. I was happy to accept your judgment, Ian, and try to learn from it.
I voted that both approaches could be appropriate. If I had a choice, I'd certainly prefer to compare wet years with wet years, year after fire with year after fire, etc. But as Chris and Cathy point out, there are many variables changing, and that difficult challenge is likely to accelerate. We need to develop techniques that allow us to evaluate change over time despite the "noise" of changing herbivores, invasives, climate, diseases, predators, successional pathways, etc. etc.
In my own conservation monitoring I've found the Floristic Quality Assessment (Gerould Wilhelm, John Taft, Doug Ladd, etc.) to be especially helpful. Previously standard measures of diversity, "evenness," etc. seemed much less useful in measuring changes in "health" from a biodiversity conservation perspective.
The basic principle of the FQA is that our conservation goal should be a high or rising diversity of conservative species. It is relatively less "distracted" by changes in rainfall, season, and other short-term quirks.
An interesting discussion. One way to address the issue is simply to recognize that two points in time are not sufficient for measuring change in ecosystems. Year to year fluctuations in management, climate, and other factors can alter some attributes of those ecosystems pretty dramatically, but measuring them at a single point in time only shows where the ecosystem is within its range of variability. The bigger question is whether that range of variability has changed - has it shrunk, expanded, or moved in one direction or another? In other words, you really need to be looking at a running average over time and looking for long-term trends.
As an example, herbaceous vegetation structure can be very different based on herbivory rates. Those herbivory rates might be driven by stocking rates of livestock or by population cycles and by rainfall - which alters the amount of total forage available. Just because the structure is short one year doesn't mean it won't be tall the next. If vegetation structure is the most important variable to assess, it's obviously not fair to assess the capacity of the ecosystem to create vegetation structure by looking at it only in a year of drought stress and/or high stocking rates of herbivores.
Excellent discussion.
Chris
I think not just weather but consider any variable that impacts on ‘condition’....and taking a step back and the notion of ‘vegetation condition’ may need unpacking. Condition for what? may be a question. Think of change, evolution (long and slow) or natural disasters (episodic and rapid) – prevailing conditions advantage some species or ecosystems, whilst wiping others out. If there is increasing variability, as we’re expecting with the impacts of climate change then we may need to revisit what we think of as ‘vegetation condition’ and how we grade it. What would you think about differences like natural seasonal variability? Could we think differently, temporal condition maybe? Which then moves my mind towards resilience theory and back loops and I’ll stop here.
My 2 cents is up.
Thanks for your postings Ian, keeps my brain cogs a-whirring.
Best
Cathy