Ash dieback continues its relentless spread through susceptible populations, and the monitoring programs tracking its progression have evolved significantly over the past few years. What we’ve learned is that different survey methodologies capture different aspects of the disease, and understanding those differences matters for forest management decisions.

Ground-based surveys remain the gold standard for assessing disease severity at individual tree level. Walking through stands and rating crown condition, documenting lesions, tracking epicormic growth—this detailed assessment provides information you simply can’t get from remote sensing. The problem is scalability. You can thoroughly survey a research plot or a managed woodland, but covering large forested areas this way isn’t practical.

The typical ground survey protocol involves rating trees on a severity scale from 0 (no visible symptoms) to 5 (dead). Seems straightforward, but there’s significant observer variation in how people assign those ratings. A tree with 30% crown dieback might be rated as a 2 by one surveyor and a 3 by another. When you’re trying to track disease progression over time, that inconsistency creates noise in the data.

Some programs have addressed this by using calibration exercises where surveyors independently rate the same trees, then compare results and discuss the differences. It improves consistency, but never eliminates subjective judgment entirely. A heavily trained surveyor with years of experience still outperforms someone using the protocol for the first time.

The timing of ground surveys matters more than many protocols acknowledge. Ash dieback symptoms are most visible in mid to late summer when crown thinning becomes apparent. Survey the same stand in early spring and you’ll underestimate severity because leaf flush can temporarily mask the extent of crown damage. Yet many monitoring programs conduct surveys when it’s convenient for field schedules rather than when symptoms are most reliably detectable.

I’ve been following data from a multi-year monitoring program that surveys the same plots quarterly. The seasonal variation in apparent disease severity is striking. Trees rated as moderately affected in August might appear minimally affected in May. The disease hasn’t regressed—you’re just seeing seasonal differences in symptom expression. Any long-term monitoring needs to account for this or ensure surveys happen at consistent phenological stages.

Another ground survey challenge is detecting early infections. By the time crown symptoms are obvious enough for a surveyor to notice from the ground, the tree has been infected for at least one growing season, possibly longer. The initial infections—the ones you’d most want to detect for management intervention—are the hardest to spot.

Some researchers have explored methods for early detection, including inspecting lower branches for characteristic lesions or looking for epicormic shoots that often appear before crown dieback becomes severe. Specialists in this space are also testing machine learning approaches to identify early-stage infections from imagery. These approaches improve early detection but require more time per tree, reducing the area you can survey with available resources.

The spatial pattern of infection within a stand tells you something about spread mechanisms. Finding diseased trees in tight clusters suggests local spore dispersal from an infection point. Finding a scattered pattern across a stand suggests windborne spores arriving from more distant sources. Ground surveys can map individual tree locations and disease status, revealing these patterns in ways that coarser survey methods can’t.

What ground surveys can’t easily do is provide landscape-level context. How is the disease progressing across an entire region? Are certain landscape positions more affected than others? Is proximity to water influencing severity? Answering these questions requires survey coverage at a scale that ground-based methods can’t efficiently provide, which is where aerial and satellite monitoring comes in as a complement.

There’s also the question of what to do with ground survey data once you have it. I’ve seen programs that collect excellent field data, enter it into spreadsheets, and then… nothing. No analysis, no integration with other data sources, no use in management decisions. The survey becomes a compliance exercise rather than a tool for informed action.

Better programs use ground survey data to calibrate and validate other monitoring approaches. Collect detailed ground data on a sample of plots, then use aerial imagery or remote sensing to extend those observations across larger areas. The ground data provides the accuracy and detail; the remote sensing provides the spatial coverage. Together they’re more useful than either alone.

One aspect that doesn’t get enough attention is surveying for resistance or tolerance. Most monitoring focuses on documenting disease severity, but the trees that remain relatively healthy in heavily infected stands are potentially valuable for understanding resistance mechanisms. Systematically tracking which trees stay healthy while their neighbors decline creates data that breeding programs need.

The cost-effectiveness of ground surveys varies hugely depending on site accessibility and vegetation structure. Open woodland with good access? You can survey efficiently. Dense understory with rough terrain and poor access? Survey time per tree increases substantially, and you may not even be able to see tree crowns clearly enough to rate them accurately.

I’m seeing more programs adopt risk-based sampling rather than trying to survey everything. Identify high-value stands or areas where early detection would enable management intervention, and concentrate ground surveys there. Use less intensive methods for broad monitoring across the rest of the landscape. That allocation of effort makes sense given limited resources.

The future of ash dieback monitoring probably involves integrating multiple data sources: ground surveys for detailed plot-level data, aerial imagery for landscape coverage, environmental data to model risk factors, and genetic data to understand resistance. No single method tells the whole story, but together they provide a more complete picture of how the disease is progressing and where management efforts might make a difference.

For anyone designing or implementing monitoring programs, the key is being clear about what questions you’re trying to answer and choosing methods that actually address those questions. Don’t survey just because “we should monitor”—survey because you have specific decisions that depend on what you’ll learn.