Gsd-skill-creator field-identification
Identifying species in the field using dichotomous keys, gestalt recognition, diagnostic features, and habitat context. Covers the working identification protocol from first encounter through confidence-rated record, the vocabulary of field marks, the discipline of negative evidence, and the honest reporting of uncertain IDs. Use when the task is to name a living organism encountered in the field.
git clone https://github.com/Tibsfox/gsd-skill-creator
T=$(mktemp -d) && git clone --depth=1 https://github.com/Tibsfox/gsd-skill-creator "$T" && mkdir -p ~/.claude/skills && cp -r "$T/examples/skills/nature-studies/field-identification" ~/.claude/skills/tibsfox-gsd-skill-creator-field-identification && rm -rf "$T"
examples/skills/nature-studies/field-identification/SKILL.mdField Identification
Field identification is the act of assigning a name to a living organism observed in its habitat, using only features that are visible, audible, or otherwise accessible without collection or dissection. It is the foundational skill of nature studies because every other activity — ecology, ethology, phenology, citizen science — depends on knowing what species you are looking at. This skill catalogs the working protocol, the vocabulary of features, the discipline of confidence reporting, and the most common ways that field IDs go wrong.
Agent affinity: peterson (diagnostic features, field-guide methodology), linnaeus (rank and naming after the field ID is made)
Concept IDs: nature-outdoor-observation, nature-plants-fungi, nature-animals-birds
The Identification Protocol
A reliable field ID is not a guess. It is the output of a short protocol that begins the moment you notice an organism and ends with a record that states both the name and the confidence.
Step 1 — Freeze the observation
Before reaching for a field guide, capture the observation. The organism may move or vanish in seconds, and memory decays fast.
- Take a photo or recording if possible.
- Note the time, location, weather, and what the organism was doing.
- Sketch the shape and any features that struck you as diagnostic.
- Note size, either in absolute units or relative to a known object.
The goal at this step is not to name the organism — it is to make sure you can still work on the identification an hour from now if the organism flies away.
Step 2 — Narrow the taxonomic frame
Before trying to name the species, name the broader group. A warbler-shaped bird with a hook-tipped bill is not a warbler. A mushroom with gills and a ring on the stalk is not a polypore. Getting the family or order right eliminates most of the lookup space.
Narrowing uses three signals:
- Gestalt. The overall shape, posture, and movement of the organism. Experienced observers can name the family in a glance because gestalt is learned, not derived.
- Field marks. Specific anatomical features that distinguish taxonomic groups. Wing bars, head stripes, leaf venation, cap texture, antenna shape.
- Behavior and habitat. What the organism was doing and where. A small brown bird creeping headfirst down a tree trunk is almost certainly a nuthatch.
Step 3 — Key to species
With the broader group fixed, work through a dichotomous key or field guide to narrow to the species. Good practice is to work from multiple features rather than leaning on one striking mark. A single field mark can be ambiguous, worn, or atypical for the individual in front of you.
A worked example for North American birds: a small gray-backed songbird with a black cap, white cheeks, and a clear two-note whistled song is a Black-capped Chickadee (or Carolina Chickadee south of the overlap zone). Three features confirm the ID: cap color, cheek color, and song pattern. Any two of these alone could be ambiguous; all three together rule out the alternatives.
Step 4 — Eliminate alternatives
Once you have a candidate, explicitly list the species it could be confused with and rule each one out. This is the single step most beginners skip. The habit of asking "what else could this be?" is what separates a reliable ID from a plausible ID.
Step 5 — Rate your confidence and record
Before writing the ID down, assign a confidence level:
| Confidence | Meaning | When to use |
|---|---|---|
| Certain | Unambiguous features, no confusion species in range | Diagnostic field marks visible, clear photo or recording, habitat matches |
| High | Very likely but one small uncertainty remains | Most features match but one is ambiguous (worn plumage, distant view) |
| Probable | Best guess but alternatives are not ruled out | Gestalt is right but field marks are incomplete |
| Tentative | Group is right, species is a guess | Know it is a warbler, cannot say which |
| Group only | Only the family or order is identifiable | Distant raptor in silhouette |
A "tentative Cooper's Hawk" entered into a phenology record is more honest and more useful than a falsely confident "Cooper's Hawk." Uncertainty is data.
The Vocabulary of Field Marks
Field marks are the specific, named features that field guides use to separate species. Knowing the vocabulary is prerequisite to using a guide efficiently.
Birds
- Head pattern: crown, supercilium (eyebrow), eyestripe, eye-ring, auricular (cheek), malar (mustache), throat, crown stripe.
- Wing pattern: wing bars, primary projection, wing edges, speculum (on ducks), coverts, scapulars.
- Tail pattern: tail band, tail spots, outer tail color, undertail coverts, tail shape (square, rounded, notched, forked).
- Underparts: breast streaking, breast band, flank color, belly wash, undertail pattern.
- Structural: bill shape and color, leg color, overall posture, tail length relative to body.
Plants
- Leaf: shape (lanceolate, ovate, palmate, pinnate, lobed), margin (entire, serrate, dentate, crenate), venation (pinnate, palmate, parallel), arrangement (alternate, opposite, whorled), texture (glabrous, pubescent, glaucous).
- Flower: symmetry (actinomorphic, zygomorphic), number of parts (petals, sepals, stamens, carpels), inflorescence type (raceme, panicle, umbel, head, spike).
- Fruit: type (berry, drupe, pome, capsule, achene, samara), dehiscence, color, size.
- Habit: tree, shrub, vine, herb, annual vs. perennial, growth form.
Fungi
- Cap: shape, texture, color, margin, scales, bruising reaction.
- Hymenium: gills (attachment, spacing, color, edge), pores (size, shape, color), teeth, ridges.
- Stipe: presence, shape, ring (annulus), volva, texture, bruising.
- Spore print color: white, pink, brown, black, purple-black, rust. Essential for many groups and cannot be determined without a print.
- Microscopic features: spore size and ornamentation (for cases where the macroscopic features are insufficient).
Insects
- Body regions: head shape, thorax markings, abdomen banding.
- Wings: number, venation, texture (membranous, leathery, hardened), coupling, resting posture.
- Appendages: antennal type (filiform, clavate, plumose), leg adaptations, mouthparts.
- Behavior and substrate: many species are ID'd by host plant as much as by appearance.
Confusion Species and Negative Evidence
Every common species has at least one look-alike. The working identifier holds a mental list of confusion species for every ID they attempt.
Common confusion categories
- Closely related species. Black-capped and Carolina Chickadees. Hairy and Downy Woodpeckers. Eastern and Western Wood-Pewees.
- Convergent mimicry. Viceroy and Monarch butterflies. Many Batesian mimics in Diptera.
- Plumage variation. Juvenile vs. adult, breeding vs. non-breeding, regional variation, leucism.
- Hybrids. Brewster's and Lawrence's warblers. Hybrid mallards. Plant hybrids in oak, willow, and other prolific genera.
- Misidentified substrate. Lichen-covered bark taken as fungal infection. Moss taken as liverwort.
Negative evidence
An ID is strengthened by ruling out alternatives, not only by matching features. If you report a Black-capped Chickadee, a useful field note might read: "Song was a clear two-note whistle, not the four-note 'fee-bee-fee-bay' of Carolina Chickadee. Wing feather edging was crisp white, not dingy as in worn Carolina Chickadees."
Negative evidence is what transforms a plausible ID into a defensible one.
When Field ID Fails
Sometimes the field ID cannot be made with the available evidence. Reliable naturalists report failure rather than guessing.
Causes of failure
- Insufficient view. The organism was too distant, too brief, or too obscured.
- Insufficient light. Dawn, dusk, deep shade, or backlight washes out color.
- Worn or damaged specimen. A faded plant, a moth with scales rubbed off, a mushroom past prime.
- Out of range. The species you expect is not recorded from this location, but the alternative you would prefer is not quite right either.
- Insufficient knowledge. You have not learned the group well enough to make the call.
What to do when ID fails
- Record what you saw at the best level of confidence you can reach. "Tentative warbler, yellow underparts, white wing bar, streaked breast — possibly a Cape May Warbler but could be a worn Magnolia Warbler."
- Note what would have resolved it. "Needed to see the throat color and hear the call."
- File the observation as open. Do not force it into a species box. Open observations are legitimate records.
- Consult others. Community review (peer-verification on iNaturalist, forum posts, field guide community threads) is the standard escalation path.
When to Use This Skill
- The user describes an organism they encountered and asks "what is this?"
- The user shares a photo or recording and wants a name.
- The user is preparing for a field trip and wants to know what they are likely to see.
- The user is reviewing an older identification and wants to verify or correct it.
- The user is building a field-guide-style artifact for a region or group.
When NOT to Use This Skill
- The user wants taxonomic reorganization (promote to the
skill).taxonomic-classification - The user wants behavioral interpretation of an organism already identified (promote to
).species-interaction-tracking - The user wants habitat or ecological context rather than a species name (promote to
).ecosystem-mapping - The user wants to make a nature journal entry rather than an ID (promote to
).nature-journaling
Decision Guidance
Begin every identification request by asking:
- Is the question really an ID question? Sometimes the user already knows the species and wants something else (behavior, range, conservation status).
- What evidence does the user actually have? A blurry phone photo is different from a clear recording is different from a sketch is different from a memory of color and behavior.
- What confidence level is achievable? If the evidence supports only "tentative warbler," do not pretend certainty. Report the group, the candidates, and what would have resolved it.
- What is the cost of a wrong ID? A phenology record in iNaturalist is low-stakes; a species report for a rare conservation target is high-stakes. Calibrate confidence thresholds to the downstream use.
Cross-References
- peterson agent: Field-guide methodology, diagnostic features, confusion species.
- linnaeus agent: Taxonomic placement after the ID is made, naming and rank.
- audubon agent: Bird-specific identification, song, flight pattern.
- taxonomic-classification skill: Placing a named species in its formal hierarchy.
- nature-journaling skill: Recording the observation as a permanent field record.
References
- Peterson, R. T. (1934, and many subsequent editions). A Field Guide to the Birds. Houghton Mifflin. (The book that introduced the Peterson System — arrows to diagnostic marks, comparative plates.)
- Sibley, D. A. (2014). The Sibley Guide to Birds, 2nd ed. Knopf.
- Kaufman, K. (2000). Kaufman Field Guide to Birds of North America. Houghton Mifflin.
- Arora, D. (1986). Mushrooms Demystified, 2nd ed. Ten Speed Press.
- Pyle, R. M. (1981). The Audubon Society Field Guide to North American Butterflies. Knopf.
- Elpel, T. J. (2013). Botany in a Day. HOPS Press.
- iNaturalist community identification guidelines (https://www.inaturalist.org/pages/help).
- eBird reviewer guidelines, various regions.