7 Reasons Satellite Images Miss Parts of Solar Shading

An overhead image can make a roof feel fully observed. You see the planes, a chimney, a tree canopy, and perhaps a neighboring building. Then the design screen adds panels, and the site begins to look settled.

Shadows move through the day and year, change with object height and vegetation, and affect electrical configurations differently. An overhead view can also hide the object that casts a shadow. A satellite image is valuable evidence, but it is one observation within a much larger model.

This guide is for solar designers, site assessors, sales engineers, and reviewers deciding how far remote evidence can carry an early layout. It does not prescribe a universal survey method. Project stage, jurisdiction, roof form, available data, and consequence should determine the verification plan.

Results depend on source data, assumptions, equipment models, configuration, and review. Outputs support design and documentation workflows but do not replace approval by the responsible engineer, authority, lender, insurer, or utility.

Treat imagery as a source with a date and purpose

Do not paste an image into the project and let it become timeless. Record the provider, capture or currency information available, resolution or quality notes, viewing limitations, and the decision it supports. If the capture date is unknown, record that absence instead of guessing from roof color or vegetation.

Google’s Solar API overview describes building and solar data products derived from imagery and elevation information. Its data-layer documentation distinguishes imagery, digital surface models, flux data, and building masks. That distinction is useful even if a team uses another provider: the photograph, elevation representation, and modeled solar data are different records with different limits.

Satellite and aerial photographs are different acquisition sources. A provider may combine photographs with measured or model-derived elevation data and solar calculations. The limitation discussed here is reliance on the visible image alone, rather than a claim that every remote-data product lacks height or seasonal information.

Use satellite or aerial imagery to support the question it can answer. “Is there enough visible roof area to justify a closer look?” may be reasonable. “What exact annual shade loss will this installed system experience?” requires more evidence and modeling.

Read the data layer before interpreting the roof

A photograph, surface model, and solar model answer different questions. Google’s Solar API data-layer reference distinguishes the following records. Other providers may use different definitions; check the documentation for the data actually used.

The reference also separates imageryDate from imageryProcessedDate. The first describes the source imagery’s approximate capture date; the second describes processing completion. Neither is the date you downloaded the file or evidence of a new site visit.

Google Research’s satellite solar-potential method describes deriving height maps and roof segments from satellite imagery using machine learning. Model-derived height is useful evidence, but it should be identified as an estimate. The researchers note limitations involving resolution, cloud cover, and occlusion; do not turn a sharp rendering or a raster pixel size into a project-wide survey-accuracy guarantee.

What can satellite imagery establish in a solar shading review?

Satellite imagery can establish a dated overhead observation of visible roof shape, surrounding context, large objects, and capture-time shadows within its resolution and viewing limits. It can support screening and obstruction questions. It cannot by itself verify current conditions, hidden features, heights, roof function, access, annual shade, electrical response, production loss, or suitability for engineering, procurement, construction, or financial commitments.

Use this copy-ready imagery evidence card:

1. Project, site, building, roof zone, and current design revision:
2. Imagery or data provider and exact layer used:
3. Capture, currency, retrieval, and processing dates available:
4. Native detail, viewing angle, orientation, crop, and quality limitations:
5. Visible roof edges, planes, large objects, surrounding context, and capture-time shadows:
6. Areas hidden by canopy, parapet, equipment, shadow, cropping, glare, or image artifacts:
7. Geometry or object fields inferred from the image and confidence assigned:
8. Other views, surface data, plans, photographs, customer statements, or measurements compared:
9. Conflicts among sources and the reviewer authorized to resolve them:
10. Permitted use, such as screen, concept, evidence request, or preliminary model:
11. Prohibited reliance and decision-changing field evidence still required:
12. Owner, expiry event, downstream outputs, and review trigger:

Separate visible evidence from inferred geometry. A crisp outline may support object location while leaving height unknown. A shadow may support that an object interrupted light at the capture moment while leaving the object’s exact geometry and annual effect unresolved. A building mask can help isolate a roof and still omit small penetrations or access constraints.

Label the permitted use precisely: roof screening, preliminary layout, or evidence request. Keep the source revision attached to derivative renderings and proposals so a new image does not silently leave old shade and energy outputs in circulation.

1. The image may predate the condition you are selling

Roofs change. Trees are pruned or grow, HVAC equipment moves, skylights appear, neighboring construction rises, and a re-roof can alter edges or penetrations. An undated image invites the team to treat a past condition as the present site.

Look for currency clues, but do not convert clues into a capture date. Compare the image with customer photographs, survey records, permit drawings where appropriate, recent street-level evidence, and the customer’s description of work. Ask directly whether the roof or surrounding site has changed.

Create a change question in intake:

If current evidence is unavailable, keep the output preliminary and name the verification trigger. “Remote concept based on imagery of unconfirmed date” is more useful than a broad note that site conditions may vary.

2. An overhead view compresses height

Shade length and direction depend on the three-dimensional relationship among the sun, receiving surface, and obstruction. A tree crown, parapet, vent, chimney, or adjacent wall may be visible from above while its height remains unknown. Other objects disappear behind taller features or blend into similar roof textures.

Some datasets include measured or model-derived elevation information. The reviewer should know which layer and method the workflow used, their limitations, and whether small roof objects are represented. Do not infer a measured height merely because the rendered model looks three-dimensional.

Check heights against field measurements, suitable survey products, drawings, known building data, or other defensible evidence. Where exact height is not necessary for screening, use a visibly labeled scenario and test whether the conclusion changes across a reasonable project-specific range. Do not publish the range as fact unless its inputs and calculation are validated.

The important distinction is simple: object location and object height are different claims. A sharp image can support the first and remain weak on the second.

3. One view cannot show every roof feature

Image angle, shadows at capture time, roof color, resolution, compression, canopy cover, and occlusion can hide small or low-contrast features. Pipes, vents, drains, wires, antennas, roof hatches, small mechanical units, and temporary objects may not read clearly from an overhead view. Parapet and setback relationships can also be hard to interpret without oblique or field evidence.

Run an obstruction sweep before placing modules:

1. Trace every roof plane and height transition.
2. Mark every visible object without deciding whether it matters yet.
3. Compare other views, photographs, drawings, and notes.
4. Create an unknown-object list with location and affected decision.
5. Route each unknown to evidence, field verification, or an explicit design exclusion.

The missed-obstruction checklist expands that sequence. The key is to inspect the source before admiring the final layout. Once modules fill the plane, reviewers tend to see the design they expect rather than the object the model omitted.

4. The sun moves while the image stays still

A shadow visible in imagery records one solar position and viewing moment. It does not show when the same object shades the array in another season, how rapidly the shadow crosses modules, or which hours carry the most solar resource.

Solar position changes with location, date, and time. The receiving plane’s orientation and slope also change how a shadow intersects the proposed array. A winter shadow may reach much farther than a summer one, while the energy consequence depends on the resource available during the affected periods.

Build a time-aware model using suitable geometry and resource data, then inspect representative dates and hours. The seasonal sun-path discussion explains how to communicate those changes without turning one animation into a production guarantee.

Avoid screenshots labeled “shade-free” unless the label defines the tested time basis and evidence. No visible shadow at noon on one date does not establish annual absence of shade.

5. Vegetation is a living assumption

Trees have height, canopy density, seasonal behavior, growth, pruning history, health, ownership, and management constraints. A satellite image can show a crown at one time. It cannot establish a future maintenance agreement or guarantee that the tree will keep the modeled form.

Separate three scenarios:

• Present observed vegetation
• Approved work supported by a responsible-party commitment
• Speculative trimming or removal discussed only as an option

Only the first two may support a project scenario, and the second needs documented scope, timing, ownership, and permissions appropriate to the site. Never improve the shade result by deleting a tree from the model because someone said it “could probably” be trimmed.

Record canopy assumptions in the same place as the shade output. If future growth is material, define a review or maintenance trigger without inventing a universal growth rate. Site species and conditions differ, and property rights can matter as much as biology.

6. Horizon and neighboring objects may sit outside the crop

Design screens naturally zoom to the roof. That crop can omit a distant tree line, adjacent tower, upper building level, terrain, pole, or neighboring structure that matters at a low solar angle. A clean roof image can therefore produce a cramped mental model of the site’s solar horizon.

Expand the review area before finalizing shade objects. Inspect multiple scales and directions. For complex sites, use field horizon measurements, suitable elevation data, or survey methods proportionate to the decision. Document which surrounding features were modeled, excluded, or could not be resolved.

The Department of Energy’s public photovoltaic design overview notes the importance of solar resource and site conditions in system design. It is educational context, not a substitute for a local obstruction survey.

When a nearby object is outside the property, also separate physical modeling from commercial control. The team can model an existing tree without implying that the customer can remove or maintain it.

7. Physical shade and electrical response are different layers

A shadow map describes where modeled direct light is obstructed. Energy effect also depends on when the event occurs, the irradiance components and resource, module behavior, bypass paths, array configuration, power electronics, and the performance model. Two layouts with similar shaded area can therefore produce different modeled outcomes.

Sandia’s PV Performance Modeling Collaborative places shading, soiling, and reflection losses within a larger modeling chain. Reviewers should preserve that chain. Do not convert a colored roof overlay directly into an annual loss statement without the intermediate assumptions and calculation.

Inspect module placement and electrical grouping together. Confirm that the equipment model and configuration match the current design revision. If a shade mitigation feature is claimed, support it with current manufacturer documentation and a model appropriate to the actual configuration. Avoid generic promises that electronics “eliminate” shade losses.

The shading analysis design workflow provides a broader handoff method. The satellite-image boundary is only the first control.

Build an evidence ladder for shading decisions

Not every project needs the same evidence, and no single method answers every question. Use an evidence ladder tied to the release stage.

Progress upward when the lower level cannot resolve a decision-changing uncertainty. Do not demand a full field exercise for a harmless exploratory question, and do not let an attractive remote model carry a construction or financial claim beyond its evidence.

Review the model before the energy number

When a proposed annual-energy value looks plausible, reviewers may stop asking how shade entered the model. Reverse that order.

First check project location, imagery and surface-data basis, roof geometry, receiving planes, modeled obstructions, vegetation treatment, horizon coverage, array placement, and electrical configuration. Then inspect resource data, diffuse and reflected components where relevant, shade calculation method, other losses, and the current equipment set. Only after those checks should the team interpret annual or monthly energy.

Keep the model name, version, inputs, loss definitions, and known limitations in the review record. These allow another reviewer to distinguish a source-data conflict from a difference in modeling methods.

If two tools differ, do not choose the preferred result. Normalize inputs where possible, document method differences, and decide which model is fit for the project purpose.

Communicate shade without turning uncertainty into fear

Customers do not need a lecture on every solar-position calculation. They need to know what was observed, what was modeled, what remains unverified, and how the uncertainty affects the choice.

Use a shade communication card:

Avoid the phrases “no shade” and “fully optimized” unless the evidence and defined scope support them. A more accurate statement often gives the buyer more control: “The preliminary model avoids the major objects visible in the current remote evidence; field verification remains required for roof features and surrounding heights.”

What should happen when remote evidence conflicts with the site?

When decision-changing imagery, drawings, or current site evidence conflict, hold the affected output and assign a reviewer to identify the controlling source. Keep both records, update the relevant geometry or assumptions, rerun dependent calculations, and issue a controlled revision. A conflict does not automatically invalidate the entire project or require one universal survey method.

Use this sequence:

1. Save the new evidence with its source, date, and affected roof zone.
2. Identify what changed, what remains fixed, and what is still unresolved.
3. Assign a responsible reviewer and the next photograph, measurement, survey, or specialist check needed.
4. Update the approved model and review affected module placement and electrical grouping.
5. Rerun affected energy, materials, and financial outputs with their own stated assumptions.
6. Issue a revision, mark prior artifacts superseded, and locate old customer or installation copies.

Physical evidence is not a direct currency claim. A new obstruction can change modeled shade and energy, while the savings effect still depends on load, tariff, export, ownership, and finance. Keep those dependencies separate. Do not attribute a changed result solely to imagery when weather, equipment, or financial assumptions changed at the same time.

Write the hold as an action: name the zone, unresolved field, affected output, evidence required, owner, and review trigger. If access is unavailable or further evidence is declined, document the resulting restriction, exclusion, deferral, or project decision. Do not fill the gap with an inferred dimension presented as measured.

Illustrative workflow example, not a customer result: Remote sources disagree about a low roof extension beside the proposed array. The reviewer holds placement and shade claims for that zone and requests suitable measurements and photographs. The unaffected screening record remains usable within its original scope.

Inspect image artifacts before interpreting objects

Stitched imagery can contain seams, displaced edges, inconsistent viewing dates, or textures that resemble roof features. Tree crowns and roof lines may lean because of viewing geometry. Bright surfaces can lose detail, while deep capture-time shadows can hide the object that cast them. These are interpretation problems before they become shading problems.

Compare multiple source layers and views where available. Zooming farther into one raster does not create new evidence after the native detail is exhausted. If the roof boundary moves between layers, record the disagreement and choose a verification action rather than averaging it by eye.

Ask reviewers to mark the source feature that supports each modeled obstruction. An object with no visible or documented basis may be an erroneous addition. A visible feature with no model object may be an omission. This one-to-one trace is especially useful on crowded commercial roofs where equipment, vents, curbs, and parapets overlap in an overhead image.

Distinguish shade avoidance from project optimization

The lowest-shade roof area is not automatically the best project location. Array decisions may also depend on usable area, structural zones, access, drainage, fire and maintenance requirements, electrical routes, inverter placement, construction sequence, roof warranty, customer scope, and consumption or export objectives.

A designer should therefore explain why the selected layout balances the documented constraints. Do not let a shade heat map become the only decision surface. Compare viable layouts on a consistent project basis and show the tradeoff that actually changes the choice.

The solar designing page describes connected roof and layout work. Any selected arrangement remains preliminary until the required site evidence and responsible reviews support its release. Shade analysis narrows choices; it does not grant permission to ignore other disciplines.

When a customer asks why a slightly shaded plane was used, answer from the controlled decision record. “Because the software chose it” is not a project rationale.

Frequently Asked Questions

Can satellite imagery be used for a preliminary solar layout?

Yes, when the imagery is suitable and the output is clearly labeled as preliminary. Record the imagery source and observation limits, identify features that remain uncertain, and state what field evidence is required before later release. Imagery can support screening, but it does not verify every roof dimension, height, obstruction, or access condition.

Why does an overhead image fail to show object height?

A visible overhead photograph does not directly measure every object height. Height can be estimated through documented remote methods or supplied by surface data, drawings, or field measurements, but the method and uncertainty must be identified. Shade duration and reach depend on those vertical relationships and the changing position of the sun.

How should trees be handled in a shading review?

Record the observation date, species or growth behavior where known, current height and canopy evidence, ownership or management constraints, and the scenario being modeled. Separate present conditions from proposed trimming or removal. Do not assume that vegetation will remain fixed or that future maintenance can occur without the responsible party’s agreement.

Does visible roof shade equal annual energy loss?

No. A visible shadow is an observation at a particular time. Annual energy effect depends on solar resource, timing, duration, geometry, array placement, equipment configuration, bypass behavior, and the model used. Review the physical shade evidence and the electrical and energy model rather than translating one image directly into an annual percentage.

When is a site visit needed for shading analysis?

Use a site visit or other suitable field method when hidden geometry, current vegetation, object heights, roof access, structural or electrical conditions, or the consequence of error cannot be resolved remotely. The required evidence depends on the project stage and decision, so define the release purpose before setting the verification level.