The short answer
Focus and light matter more than composition. An identifiable observation fills the frame with the subject, is in sharp focus, and shows the specific features an expert actually needs, leaf undersides, wing patterns, the shape of a bird's bill, not just a pretty overall shot. A blurry, distant photo of a rare species is worth less to science than a sharp, close photo of a common one. Sound recordings count too, and for some species they're the better form of evidence.
General principles that apply to almost everything
- Fill the frame. A speck in the distance is hard to identify. Get close, or zoom in. You can also crop your image to increase the focus on your subject organism, just be careful not to crop so tightly that the image gets blurry or pixelated.
- Sharp focus beats a nice background. A blurry photo is often impossible to identify no matter how good the composition is.
- Keep the sun behind you. Backlit subjects lose detail and color accuracy. Full sun or full shade is easier to work with than dappled light, which distorts colors and patterns.
- Take more than one photo, from more than one angle. A single perfect shot is rarely enough. Multiple angles showing different features give identifiers far more to work with.
- Always enter at least a rough identification. Even just "plant," "bird," or "insect" instead of leaving it blank. It helps specialists find your observation and dramatically increases the odds someone follows up.
What to photograph, by group
- The whole plant, to show its overall shape and growth pattern
- Both sides of the leaves, and how leaves attach to the stem
- Flowers from straight-on and from the side, including the reproductive parts
- Fruit or seed, if present, since a botanist often needs it to confirm identification
- Something for scale, a coin, a hand, or a shoe
- Photos from the side, top, and front if you can get them
- Close-ups of diagnostic features: wings, eyes, legs, or scale patterns
- For moths and butterflies, both the top and underside of the wings, since the two views often look completely different
- A photo of the habitat or the plant an insect was found on, since many species have strong plant associations
- Approach slowly and quietly rather than rushing, most animals give you more time than you'd expect if you don't startle them
- Top, bottom, and side views
- What it's growing on, tree trunk, soil, leaf litter, dead wood, since substrate is often a key identifying clue
- The surrounding habitat
When to record sound instead
iNaturalist accepts audio recordings as evidence, not just photos, and for some species a recording is genuinely more useful than a picture. Birds, frogs, and many insects are identified primarily by call, and a clear recording of a calling frog or singing bird can nail an identification that a distant, blurry photo never could. If you hear something you can't get a clear photo of, record it instead of giving up on the observation entirely.
Use the AI suggestions, but don't rely on them blindly
When you upload an observation, iNaturalist's computer vision model suggests possible identifications, ranked by confidence, based on the photo itself and observations already confirmed nearby. It's a genuinely useful starting point, especially for beginners, and it's often right. But treat it as a suggestion, not a verdict. The percentage shown is a confidence score based on visual similarity, not a guarantee, and it can be thrown off by unusual angles, juvenile individuals, or species that look alike. Pick the suggestion that actually matches what you observed, not just the top-ranked one, and let the wider iNaturalist community confirm or correct it from there. That back-and-forth verification is what makes the platform's data trustworthy in the first place.
A few things that trip people up
- Don't touch or pick anything you can't confidently identify. This protects both you and the organism.
- Mark cultivated or captive organisms as such. A garden plant or a zoo animal isn't a wild observation, and mislabeling it undermines the data other people rely on.
- One organism per observation, generally. Photographing multiple individuals in one observation makes identification and record-keeping harder, not easier. If there's a tight clump of the same species, that's usually fine as one observation; otherwise, separate observations are better.
- Light editing is fine, heavy editing isn't. Increasing contrast, brightness, or sharpness to make a feature more visible is generally encouraged. Avoid editing colors, since that can actively mislead an identification.
iNaturalist automatically hides the precise location of species with an at-risk conservation status, a feature called taxon geoprivacy, so this usually happens without you needing to do anything. You can also manually obscure a location yourself, useful for observations on private property, for example. What you should never do is enter a false location, even with good intentions. Accurate location data is what makes an observation useful for research and conservation in the first place; obscuring protects a sensitive species without breaking that.
Frequently asked questions
- iNaturalist, "Tips for Making Great Observations." inaturalist.org
- iNaturalist, "Checklists for Making Better iNaturalist Observations." inaturalist.org
- iNaturalist, "Creating High Quality iNaturalist Observations." inaturalist.org
- iNaturalist Help, "What is geoprivacy? What does it mean for an observation to be obscured?" help.inaturalist.org
- iNaturalist, "Geoprivacy." inaturalist.org