Evaluating Forest Inventory Technology for Small Landowners

The following post originally appeared in the 2017 issue of The Consultant. It was written by SilviaTerra's co-founder, Max Nova.

A consulting forester must wear many hats: cartographer, silviculturist, cruiser, entomologist, planner, harvest supervisor, accountant. For each of these roles, there is a constant stream of new technologies. Which of these is worth adopting? Which will best help private landowners understand and relate to their land? Given the many options, how can you decide among them?

The many facets of our profession require that we be sparing with our focus in any one - and that we make use of efficiencies that can help us collect data and relay information in meaningful ways. That is the promise of technology: to save time and improve outcomes. At their best, new tools improve our ability to understand our forests and direct their growth. At their worst, new “technologies” can increase confusion, complicate processes, increase costs, and generally distract us from what is really important.

When it comes to technology for connecting private landowners with their land, it is important to remember both the heart and the mind. Heart is the emotional connection to the land - the story a landowner tells their children about their property. Mind is the detail-oriented, analytical focus needed to create and continually revise an optimized management plan for achieving the landowner's most important objectives. Any technology can be evaluated by how well it helps landowners in these two areas.

In this article, we will lay out a framework for evaluating which technologies can help you empower landowners to connect to their properties with both heart and mind.

HEART
Small landowners have a personal, emotional connection to their land. There are often decades of stories and shared history on a particular tract. Consultants can use technology to help landowners explore their property through new perspectives, deepening and enhancing their bond with their land.

dji-phantom-4
The DJI Phantom 4 Advanced is a professional-grade drone that retails for $1,200

No article about new forest technology would be complete without a section on drones. In the past few years, several user-friendly $1,000 drones have emerged that make it possible for even small consulting shops to own and operate their own drone. Drones give landowners a new point of view, creating a powerful visual tool for a consultant explaining proposed management plans. Drone imagery particularly shines as a communication tool because clear, high-resolution images of a single property lend a sense of immediacy and reality to a discussion.

The typical scale of consulting forestry is well-suited to drone flights. Drones have limited utility in larger, industrial-scale forests because of restrictions on flight ceiling, battery life, and line-of-sight operation. Because most consulting work is done on a parcel-by-parcel basis, consulting foresters neatly avoid most of these logistical issues.

But drones are not the only way to get imagery for a property. Many landowners already know how to use Google Maps to view recent satellite imagery for their tract but, few know of an even more powerful tool: Google Earth. Consulting foresters can use Google Earth to zoom to a landowner's parcel and then drag a timeline slider to view historic satellite imagery. This is a visually compelling way to discuss past land-use changes on and around the property. There is nothing like being able to see a series of photos to tell a story about the silvicultural history of a particular stand and how that history plays into future management possibilities.

inaturalist--1-
The iNaturalist app is great for getting the younger generation engaged

Technology has more to offer the heart of the landowner than just pretty pictures; it can also help get younger generations involved in the land. We frequently hear that older landowners struggle to get their children and grandchildren interested in their property. Technology offers a way to bridge the gap by letting the "digital native" generation connect with the property on their own terms. One of the best examples of this is the iNaturalist app. Developed by the California Academy of Sciences, iNaturalist is an iOS/Android app for recording and identifying plant, insect, and animal species. Using the app, you take a picture of an organism. The app records the date and location and saves the picture to your account. With multiple observations, you can begin to build up a record of the species diversity and the various habitats present on the property. You can even become a "citizen scientist" by sharing your observations with the scientific community. And for unknown species, you can share the picture with experts on iNaturalist to get probable identification. Apps like these provide a highly visual, hands-on way for getting the next generation involved with the land and can spark intergenerational conversations about forest management and the future.

MIND
A landowner's emotional connection with the land is a key driver in determining the values and objectives of a management plan. But rigorous and efficient long-term planning must include quantitative analysis as well. The mind gets involved when landowners want to optimize the management of their property to achieve their values. When is a new technology a cost-effective way to improve management outcomes? As consulting foresters, how can we show that a management plan is truly optimal?

This is a particularly difficult question to answer in an analytically rigorous way. There are so many interconnected factors involved in forest management that it is hard to quantify the impact that each decision has on the ultimate outcome. We can use the question, "How many plots should I put in this stand?" as an example to show how you can put some hard numbers to deceptively complex questions like this. There is a detailed explanation of this approach on the SilviaTerra blog, but we will sketch out the top-level view here.

Having good inventory information is critical for being able to develop a good management plan. If the underlying inventory data is off, a management plan might recommend suboptimal treatments - harvesting too early, too late, etc. No forester has the time to conduct a census on each parcel, so we always accept that our inventories (and thus our management plans) are going to be a bit off. Most of us have rules of thumb that we use to decide on how many plots to install in a given stand - maybe 1 plot every 3 acres in pine and 1 plot every acre in hardwood. But is that actually optimal? Are we cruising too little relative to the value of the decisions we are making based on that data? Or are we spending too much on cruising?

forestMap
A "virtual forest" created using the USFS FIA dataset

Here is how we can know. Let us consider a 200 acre, 10 stand ownership in Northern Georgia that is predominantly 12 year old loblolly pine. We can use the excellent, spatially explicit USFS FIA dataset to select a plot from a similar forest type and generate a "virtual forest" where we know the location, diameter, species, etc. of every tree. Because we have perfect information, we can establish a baseline for what a perfectly optimal management plan would look like and how much revenue it would generate. Then we simulate "cruising" our virtual forest. We can put in a BAF 10 plot every 3 acres and see what inventory estimate comes out. Then we build a management plan based on that estimate and see how we do. Because we will be working off of imperfect information, it is likely that we will be cutting some stands too early and others too late.

But now we can quantify exactly how much money has been lost compared to having perfect information. Adding in the per-plot cost of data collection, we can get a sense for the overall economics of our sampling strategy. We can try many other sampling intensities and methods to see how they perform. At the end of our process, we see which strategy has the lowest cruising cost + expected management loss, and that is our optimum.

This process can be run for many different sampling strategies and in many different forest types, and can also be adapted for non-commercial objectives like creating species habitat or other ecosystem services. An example of this type of analysis can be found on the SilviaTerra website.

This approach also points towards a way of evaluating new technologies like recent developments in remote sensing. Most remote sensing forest inventory technologies work by reducing the number of plots that we have to cruise. Many foresters already stratify their forests to reduce the variability within each strata - remote sensing extends this approach by working at a finer resolution and by applying statistical models to identify areas that are similar to each other. If the cost of the technology is not too great, this enables us to get better data and make better decisions while coming out ahead financially. In general, we should consider adopting a new technology if the improvement in our management outcomes outweighs the cost of implementing the technology.

Over the past few years, there have been several exciting advances in remote sensing platforms, sensors, and software that are opening up new possibilities for forest managers. However, there are several structural barriers that make it difficult for consulting foresters to adopt some of the latest remote sensing technology - primarily high fixed costs and technological expertise. Which of the available options makes sense for most consulting foresters?

One of the buzziest technologies out there now is called LiDAR - short for "light detection and ranging". It's essentially a rapidfire laser rangefinder on a swivel. The basic idea is that you strap a LiDAR sensor onto the bottom of a plane, it takes really high resolution scans of the trees below, and then you analyze the resulting "point cloud" to develop a forest inventory. There are a couple of problems with this for most forestry consultants. The first is cost - a LiDAR scanner costs tens or even hundreds of thousands of dollars, to say nothing of the plane or pilot! The second is that converting a point cloud to a forest inventory is far from a solved problem - particularly in closed-canopy, mixed forests. It is simple and straightforward to read off average heights, but there are no readily available off-the-shelf tools for deriving accurate and unbiased forest inventories from point clouds, and most consultants do not have the time required to learn the advanced statistics and computer programming to take a crack at it. While the cost of LiDAR sensors is beginning to drop and new drone and satellite LiDAR platforms are being developed, at this point, the cost of the equipment and experimental nature of the analysis outweigh the potential benefit of incorporating LiDAR data into a consulting forester's inventory strategy.

We previously discussed how drone imagery can help catalyze conversations with landowners. The widespread consumer adoption of drones has driven their cost down while spurring the development of user-friendly software for stitching together drone imagery. This makes drones a compelling option for foresters looking to conduct quick, qualitative assessments of stands. Some researchers have even been able to develop point clouds from drone imagery by using a technique called "structure from motion." Using a sequence of images taken from slightly different positions as the drone flies, it is possible to triangulate the points of the tree canopies and develop a model of the canopy's surface. This is similar to combining lots of stereo pair images to measure height difference. However, when it comes to the development of a statistically unbiased tree-segmentation based forest inventory, drone imagery faces many of the same challenges as LiDAR point clouds. We suspect that drones will soon become a part of every foresters toolbox for rapid, qualitative stand assessment, but the days of totally remote drone-based forest inventory are still a long way off.

Microsatellite companies like Planet Labs are promising global imagery coverage with daily revisits

Finally, there have been some recent developments in satellite platforms and analysis that may be useful for consulting foresters. Of course, satellite imagery has a long history in forestry - researchers have been mapping land cover change and stratifying forests with Landsat for decades. And many of the images used in programs like Google Maps come from satellites. However, one of the challenges of using satellite imagery is that it is sometimes difficult to acquire a cloud-free image - particularly in the Pacific Northwest. A new breed of satellite is helping to solve this problem though. Companies like Planet Labs have launched fleets of "microsatellites" that are close to providing daily imagery of the entire globe. Unlike conventional, bus-sized satellites like Landsat, these microsatellites are only about the size of a few iPhones strapped together. And rather than launching just one, the microsatellite companies launch upwards of 50 microsatellites at a time. The data from these companies is generally not available to consulting foresters yet, but competition is heating up and prices should be rapidly coming down.

1) A conventional grid of plots (green) in a stand in the Pacific Northwest yielded an estimate of basal area with an error of +/- 17% at 90% confidence
2) Remotely sensed imagery adds a lot more information about the forest and helps fill in the gaps between plots
3) Total basal area distribution across the stand (redder means higher BA). Statistical analysis of a stack of remotely sensed imagery tightened up the total BA confidence interval to +/- 10%
4) Western Hemlock BA
5) Douglas Fir BA
6) Noble Fir BA.
Aerial imagery courtesy of Bing Maps

There are several options for developing a forest inventory from satellite imagery and cruise data. The simplest is to use the satellite imagery to delineate stand bounds, stratify the stands, and then work up the strata-level inventories using well-understood statistical methods taught in most introductory biometrics classes. To develop stand-level and sub-stand level inventories requires more advanced statistical methods to ensure an unbiased estimate. Currently SilviaTerra's CruiseBoost service is the only turnkey service for consulting foresters to integrate satellite imagery into their inventory calculations.

Consulting foresters hold a unique place in American forestry because of their role in helping manage small private lands. They translate the complexities of forest development into the values of the landowner - making the forest comprehensible to non-experts. Unlike institutional players like TIMOs and REITs, small private landowners are not fiduciarily obligated to maximize economic returns. They often have a longer term, smaller scale, and more emotional connection to their land. By adapting the latest practices and technologies to this more intimate context, consulting foresters serve as an essential bridge between small private landowners and the best that modern forest science has to offer.

The profusion of platforms, sensors, and software means that it has never been a more exciting time to be a consulting forester. In the next few years, what if every private landowner in the US had an updated, accurate inventory and a data-driven and optimized management plan? This is an enormous dream and it is going to take a lot of work from all of us in the forestry community. But thanks to recent developments in technology, the dream is edging towards becoming a reality. We are looking forward to continuing the conversation about technology in forestry with the ACF community.

Max Nova

Max Nova

Max is one of the founders of SilviaTerra. When he's not building new tools for foresters, Max can usually be found reading a good book or riding his bike. Max is a graduate of Yale University.

Read More