Forest Inventory — Here is the Thing in British Columbia
My wife rearranged the kitchen! What does that have to do with Forest Inventory you say? Well quite a lot. I spent 5 minutes just looking for the knives and forks so I could set the table before our guests arrive. Later, I thought she got rid of my favourite coffee cup, but no, just buried in the back and now mixed in with the water bottles.
My wife is a good woman, but all of this is to make a point. You see we all carry around a household inventory in our heads, with a built in geo-referencing system.
When I run out of coffee beans, I go to the Costco cupboard and start looking on the second shelf. It may not be in exactly the same place every time but I can count on finding it in the general vicinity. When I pull out the second to last bag of beans, I make a note on the groceries list so that we remember to get more before we run out.
Inventory is like that, or at least it should be. It is intended to keep the necessary supplies available and easily accessible at all times. Systems are put in place to ensure that we don’t run out, especially when it comes to the coffee — those supplies really need to be maintained.
…we all carry around a household inventory in our heads, with a built in geo-referencing system.
So I ask, how does this compare with forest inventory in British Columbia?
To start with, forest inventory is knowing something about trees — how many, what kind (species that is), how big, where, in juxtaposition to what (streams, lakes, roads, land forms, and many other features). More generally, it encompasses a much broader set of attributes. Still we are talking about forest inventory. We use this information to help maintain and protect those things that are really dear to us (Scottish Joke: “That is a Deer Fence Over There!”), and to make use of those that we need for another purpose.
Forest inventories beyond our family woodlot are generally too big to hold in our heads, so instead we start by making maps of where things are, what kind, how much, etc. To start with, we might do this by delineating stands, and then visit every one of them. Then we establish a few ground plots, and produce a rough estimate of attributes for each one.
As the inventory area gets bigger, we stop doing that. Instead, we stratify the stand types using remote sensing data (traditionally using aerial photographs and manual photo interpretation). Once again we can put in ground samples to estimate attributes, but this time it is for each strata instead of each stand.
As the area gets bigger, and the forest encompasses even more variation we may just start estimating the attributes, for example by photo interpretation, and hope that the numbers are close to the real thing. We might still put in plots in an effort to make adjustments for bias. To a large extent this last state of affairs represents forest inventory in British Columbia (BC).
Is this “The Best” we can do with what we have got?
I don’t think so.
Lets consider first how we use inventory data of the kind just described.
I am a Timber Baron and looking for wood in all the right places. I start by looking at the map just to the west of Kelowna on the Interior Plateau. I am looking for a Spruce-Pine-Fir stand, 20 m or more in height, with 300 m3/ha or more and with “a good average piece size” (i.e. average volume per tree). I am already asking too much of the inventory. The inventory has estimates, but there is a lot of error associated with these estimates when it comes to a given stand or polygon. There can also be a lot of “within polygon” variation, such that the attribution is not all that meaningful. I use this information because I have to start somewhere. In this case we are doing the best with what we have.
Back to the analogy. It is a bit like knowing that there are coffee cups to be found somewhere in a given neighbourhood. We can pick a neighbourhood, and then a house to start with. Next, find the kitchen, mess around in the cupboards, and eventually we find something that may not be exactly what we want, but will be good enough. We find the water (tap is easy to find), the pot (French press this time), the filters (oh ya, we don’t need them), the coffee (already ground, not my favourite but pretty desperate by now), and two cups. The rest is easy, heat up some water (had to find the matches; it was a gas stove), put the coffee and water into the French Press, wait 5 minutes, and voila. Well that was easy. It took us an hour and half compared to the usual 15 minutes at home, but we found it and managed to make ourselves some coffee. Victory is sweet (darn, forgot the sugar).
Well —this is what we do in forestry in many parts of British Columbia! Using the inventory of the aforementioned kind, we pick a general area on the map that seemingly has good prospects, get in a helicopter, fly around, draw lines on the map or even better on an aerial photo as to where we think we can harvest. Later we visit a subset of these things on the ground to confirm whether or not they are good places for harvesting trees. Finally, we put in ground plots and determine the final boundaries that we think might work.
We run a whole industry this way. Uncertainties rebound in the form of additional costs of production. At a strategic level we run the risk of assuming that forest practices are sustainable when in fact they are not, or we run the risk of under-utilization with respect to any one of a number of forest uses.
In any event, the uncertainties stand as a barrier to adequately reconciling operational outcomes with strategic level outcomes. Uncertainties also lead us toward increasing log yard inventories so as to maintain profitability in the milling of timber for specific product lines.
We even invest in things like reducing fire risk around urban areas. However, we have no effective means of reflecting what was left behind in the inventory, (we didn’t really know what was there in the first place until we completed ground surveys), and as a consequence we don’t know how those practices are expected to reverberate with time.
Here is the Thing
In British Columbia, we are keeping the cost of forest inventory within certain parameters- lets say $1.50 per hectare. The money we are saving, unfortunately, is more than offset by the money we are spending (i.e. lose) to compensate for the uncertainties that exist. The alternative is to spend more on forest inventory, and in so doing, increase its reliability, level of resolution, level of detail, and flexibility of (multi-scale) interpretation.
The technology is already here: Advances in computer data storage and processing speeds following Moore’s Law, in Airborne Laser Scanning (ALS)- also referred to as Light Detection And Ranging (LiDAR), the use of Geographic Positioning Systems (GPS), the advent of Inertial Measurement Units (IMU’s), digital Colour Infrared (CIR) photography and/or satellite imagery, and the use of a standard geographic coordinate system (e.g. the Universal TransMercator system or UTM).
The opportunity is to put more emphasis (money) into establishing and re-measuring ground plots for the purpose of building and maintaining inventory. We can project changes in ground plot conditions, through the use of growth and yield models, and then compare these with actual conditions on the ground. Using the plots and all this advanced technology, along with a healthy dose of advanced data analytics, we can better develop and maintain inventories with time. In so doing, we can also reduce the inventory depreciation rate and amortize the ongoing investments over a longer periods of time.
Of course, we will have created greater value by reducing the uncertainty around the initial conditions, and thus realize greater savings and profitability as our Return On Investment (ROI). We can also reduce the risk of managing our forests into a cul de sac that may leave few options for recovery.
Ultimately, we can use ground plots as means of tracking changes in forest characteristics. This in turn may provide us insights into the impacts of climate change. Think of it as an investment in the “Canary in the Coal Mine”. The sensors (the trees) are there; all we need to do is to put some energy into picking up the signals. At that point we can better explore the potential impacts across the landscape as a whole.
What I don’t understand is why we keep finding excuses as to why we should not invest in more reliable, higher resolution inventory, starting today. It is not the technology. Surely it is not the money — i.e., the incremental cost of producing (and using) the inventory given the latest technology.
Maybe we need to do a better job of assigning a value to forest inventory information. We know in principle that it is equal to the cost of the inventory plus the additional amount paid to compensate for the (lack of) quality of the information once we get it, plus the amount lost as a result of missing the intended mark as far as strategic level outcomes are concerned. If that is the excuse, then we should fix it.
Here is the Big Thing
We should be moving ahead with a more substantive initiative to advance the quality of forest inventory information and forest inventory applications in the Province of British Columbia. We know how to do it. We have the resources to do it. Many people I talk to even suggest that we should in fact be doing it. What is lacking is the leadership to get on with doing the job. After that we can get the job done!
Read more about Tesera’s approach to high resolution forest inventory.
Ian Moss, RPF is Chief Analytics Officer at Tesera.com
