Solar Panel Tilt Angle and Azimuth Guide 2026

What Tilt and Azimuth Optimisation Actually Means on a Live Design

Set tilt equal to site latitude, face true south in the Northern Hemisphere or true north in the Southern, and you will be within a percent or two of maximum annual yield for most of the year — that is the rule every generic guide repeats. It is a reasonable starting point for a mid-latitude site and it is where most quoting tools default. But it is a rule of thumb, not a result, and it is not where the industry's own modelling data lands once you go looking for the number behind it.

The most rigorous public source for country-level optimal tilt is Jacobson and Jadhav's 2018 paper in Solar Energy. Jacobson and Jadhav's 2018 paper in Solar Energy built its country-by-country optimal tilt table by running NREL's PVWatts program, rather than by applying a latitude formula. That is an important distinction for a design team: the table did not derive tilt from latitude at all, it derived it from simulated output and then compared the result to latitude afterward. When the authors ran that comparison, they found the two diverge sharply at higher latitudes. A third-order polynomial fit of optimal tilt against latitude matches that PVWatts data better above 40 degrees north than the straight-line latitude rule does. In plain terms: the latitude rule is roughly right in the temperate band most EPCs quote in, and increasingly wrong as you move toward the pole.

Vendor-published guides independently arrive at the same shape: optimal tilt runs 5–15 degrees below latitude for most locations, because summer months carry most of the year's irradiance. A flatter-than-latitude tilt trades away some low winter-sun capture in exchange for catching more of the high-sun summer months when the site has the most irradiance to gain from — and since summer dominates the annual total, the trade nets positive.

None of this means the PVWatts-derived numbers are what an EPC should build to unmodified, either. Jacobson and Jadhav caution, in their own words, that their optimal tilt values are not necessarily the most cost-effective fixed tilt angles, because the dataset excludes the extra land needed to keep rows from shading each other. That caveat is the hinge the rest of this guide turns on: energy-optimal tilt and cost-optimal tilt are different numbers, and on any site where land or roof area is the binding constraint, the second number wins.

The Four Decisions That Actually Set Your Numbers

The Latitude Rule and Where It Breaks

Treat tilt = latitude as a default for mid-latitude sites, and treat it as a starting point that needs correcting at higher latitudes. Above 40 degrees north the PVWatts-derived data shows the true optimum bending flatter than a straight latitude line predicts, and guidance of 5–15 degrees below latitude gives a workable adjustment range to apply before running a full simulation. Near the equator the logic reverses for a different reason entirely: Sunlith Energy's guide sets a minimum tilt of 10–15 degrees at 0–15 degrees latitude — not for irradiance capture but so rain can self-clean the panels, since near-flat panels accumulate dust and pooled water. At the equator, in other words, the tilt argument isn't about sun angle at all.

How Much a Miss Actually Costs

This is the number an EPC actually needs before deciding whether a non-ideal roof is worth quoting. Publications' guide states that deviations of up to 15 degrees in tilt and 30 degrees in azimuth cost under 5% of yield. Push further and the cost climbs faster: azimuth table, built from central European (45–52 degrees north) data, gives 76–80% of south-facing annual yield for a due east or due west array — a 20–24% loss at that latitude band. That is a real number to walk a client through before quoting an east- or west-only roof, not a reason to automatically decline it.

The penalty also depends on where the site sits. Sunlith Energy's guide puts a 45-degree azimuth deviation at temperate latitudes (35–55 degrees) at a 5–8% annual yield cost, materially more than the same turn nearer the equator — the same rooftop orientation argument plays out differently in Berlin than it does near the tropics. Worth flagging to a design team: the industry sources do not fully agree on the exact percentage for a given deviation, which is itself useful information — treat any single vendor's loss table as directional, not as a number to defend to the decimal in a proposal.

Tilt misses behave the same way, own simulation work puts a number on it. It is explicit that this quick estimate is a sanity check rather than a deliverable. Anything that goes in front of a client should come from the full simulation run, not the quick orientation graph.

When Density Beats the Optimum

This is the section the generic tilt articles skip, and it is the one that actually governs commercial roof and ground-mount bids. On any site where the constraint is area rather than irradiance — a flat commercial roof, a density-limited ground plot — the energy-optimal tilt is close to irrelevant, because raising tilt costs installable capacity faster than it buys back yield per panel.

The tilt that maximises transposition is around 36 degrees, but the tilt that maximises real yield after shading losses is about 31 degrees — shading pulls the real answer flatter than the plane-of-array optimum by roughly 5 degrees before land is even considered. Documentation states shading losses increase significantly as tilt increases, because ground coverage ratio strongly decreases at higher tilts. At 30 degrees of tilt, publications state that you can install less than half the available ground area in collectors if you want to keep the same limit angle. That single fact is the real reason tilt gets set low on density-constrained sites: it isn't about chasing marginal yield, it's about how many panels physically fit.

It is advised going as horizontal as is reasonable for rain cleaning: at 5 degrees instead of 30 degrees you can install almost twice the PV power on the same area, losing about 9% of per-panel efficiency but gaining more in total energy. That is the trade an EPC needs to run for every area-constrained bid: total site output, not per-panel efficiency, is what a flat-roof or ground-mount client is actually buying. The row-spacing side of that trade has its own guardrail — recommend keeping the limit angle to 18–22 degrees, because a higher ground coverage ratio produces higher mutual shading, especially electrical mismatch. Pack rows tighter than that and the mismatch losses start eating the density gain.

Going low-tilt has a module-spec consequence worth putting in a proposal note: at low tilts frameless laminates should be used, because a framed module traps dirt and moss along its bottom rail. And flattening the array has a second, quieter benefit on non-south buildings: lowering shed tilt also lowers how much yield depends on azimuth. A low-tilt array sidesteps that choice almost entirely.

There is no single answer to shed optimisation because it is a multi-variable, multi-criteria problem — available area, investment, support cost, weight limits and power limits all pull toward different tilts depending on whether you are optimising energy, system efficiency or cost of energy. That is the honest answer a design team should give a client who wants a single "best" number: it depends on what the client is actually paying for.

Fixed Tilt or Tracker

Once tilt is set, the racking decision follows from the same density-versus-output logic, now applied across the whole plant rather than within a single row. Mobile racking passed 50% of US utility-scale projects starting in 2017, according to the U.S. Energy Information Administration, was installed in more than half of U.S. utility-scale solar projects starting in 2017, and the manufacturing base backs that up — U.S. manufacturers produced 37 GWdc of solar trackers in 2023, a year in which the country built 32.4 GWdc of new solar overall.

But fixed tilt has not been displaced everywhere, and the reasons are specific rather than nostalgic. Fixed-tilt racking remains the better choice in high winds, heavy snow loads and steep topographies, according to Solar Power World's April 2025 reporting. The wind case is the sharpest: as OMCO Solar's director of engineering told Solar Power World, in high-wind conditions trackers stow flat, so the array is effectively running as fixed tilt and the tracking benefit is lost. A site that stows often enough is paying tracker capex for fixed-tilt output. Slope is the second case — fixed-tilt units have higher slope tolerance than trackers, which cuts the grading a site needs before construction, since short fixed-tilt bays can follow rolling ground where a long tracker torque tube cannot.

Density repeats the shed-tilt trade at plant scale: fixed-tilt can be installed at higher system density than trackers, because tracker arrays have to be spaced to avoid inter-row shading. And the cost side is straightforward — fixed-tilt racking also costs less to manufacture, ships in fewer truckloads, and needs less specialised install labour than tracker hardware.

Where the fixed-versus-tracker line gets genuinely close is single-axis versus dual-axis tracking, and here the data argues against paying for the second axis in most cases. Jacobson and Jadhav report that 1-axis horizontal tracking comes within 1–3% of 2-axis tracking, annually averaged, at most latitudes, and that a second axis buys little because 2-axis systems also need more land to avoid shading rows behind the front row. At extreme latitude the case for any tracker weakens further: tracking of any kind provides little benefit over a well-chosen fixed tilt above 75 degrees north or 60 degrees south.

The Shading Study Calendar Every Fixed-Tilt Design Has to Clear

Row spacing and setback decisions are not made against an average day — they are made against the worst day, and the industry has settled on the same two calendar dates to test it:

• December 21 (Northern Hemisphere winter solstice) — Experienced designers use December 21 as the shading-study benchmark: a panel that is shade-free at noon on the winter solstice will be shade-free all year, and row spacing, setbacks and obstruction clearances are all derived from that sun angle.
• June 21 and December 21 together — shadow-analysis report uses June 21 and December 21 as its two study dates, bracketing the high-sun and low-sun extremes of the year rather than relying on the winter case alone.
• True south, verified against a compass reading — before either date matters, azimuth has to be set correctly in the first place. Publications' warns that magnetic declination can shift a compass reading by 1–15 degrees or more, so azimuth has to be set from true south, not a handheld compass. Get this step wrong and every downstream shading calculation for both solstice dates is off by the same amount.
• The pre-design-freeze site simulation — before any of these numbers go into a proposal, Sunlith Energy's guide advises running a site-specific simulation in PVGIS or PVWatts to account for local shading, horizon obstructions and microclimate before finalising a design, echoing Jacobson and Jadhav's own instruction that installers should calculate optimal tilt for their location. Treat every figure in this guide as the input to that run, not a substitute for it.

What EPCs Must Do Now

• Correct the latitude default before quoting anything above 40 degrees latitude. Start from tilt = latitude at mid-latitude sites; north of that, shift flatter per vendor guidance (roughly 5–15 degrees) as a first pass and confirm with a full simulation, not the quick orientation graph.
• Quote non-south roofs using the loss table, not a refusal. A 15-degree tilt miss or 30-degree azimuth miss costs under 5%; a full east or west orientation costs 20–24%. Both are quotable — the second just needs the client to see the number.
• Run the ground coverage ratio math before setting tilt on any area-constrained site. On a flat commercial roof or a density-limited plot, check what tilt the plot's economics actually support before defaulting to an energy-optimal angle that halves your installable capacity.
• Match the racking choice to site conditions, not to what utility-scale projects default to. High wind, heavy snow, steep grade, or a density-constrained plot are each independent reasons to price fixed tilt over a tracker; a single-axis tracker, not a dual-axis one, is the right upgrade when tracking is justified at all.
• Set azimuth from true coordinates, verify against the winter solstice, and never publish the quick-tool number. Confirm true south (or true north south of the equator) before drawing the layout, check both solstices for shading, and treat any fast in-app estimate as a check on the full simulation, not the deliverable.

Setting Tilt and Azimuth Inside the Design Tools

Design software treats a flat roof and a pitched roof as fundamentally different inputs, and getting that distinction backwards is a common source of layout errors. In design tools, polygon tilt is set to 0 for flat roofs, because that field describes the roof surface rather than the array itself — the actual array tilt gets set separately on the mounting structure. It is recommended to set azimuth for tilted roofs only after the polygon has been drawn, and setting subarray azimuth only after the modules are placed on it, since a default entered before drawing tends to be wrong for every plane after the first one.

On a pitched roof the decision disappears entirely in most tools: for flush-mount structures, tilt and azimuth are calculated automatically from the roof polygon rather than entered by the designer — the tilt argument only really exists on flat roofs and ground mounts. That is worth stating plainly to a junior designer: the tilt-and-azimuth optimisation this whole guide covers is a flat-roof and ground-mount problem. On a pitched residential or industrial roof, the roof has already made the decision.

Flat-roof layout strategy has one more option worth pricing before defaulting to south-facing rows: east-west flat-roof layouts, using two low-angle panel faces, practically eliminate inter-row shading and fit 30–50% more panel capacity on the same roof area. On a roof where south orientation isn't available or row spacing is eating too much of the plot, an east-west low-tilt layout is a real alternative to quote, not just a fallback.

Finally, every number in this guide sits on top of a weather-year assumption that is worth stating in any proposal that promises a specific annual output. In the US, solar output in a 10th-percentile weather year runs on average 4.8% below a 50th-percentile year — a swing bigger than most tilt or azimuth arguments being had. That variability exists because the underlying tool is built on historical averages rather than a forecast: PVWatts itself is built by combining site solar resource data with 30 years of historic temperature and wind speed data from a nearby weather station, which is why its output is an approximation for any single site. For a citable, published reference behind photovoltaic performance modelling, NREL's technical reference for the performance model inside the System Advisor Model, report NREL/TP-6A20-64102, is the free, citable source to point to.

Frequently Asked Questions

Q1. Does tilt = latitude still hold as a rule for 2026 designs?

It holds reasonably well at mid-latitude sites, where NREL PVWatts-derived data and the straight-line latitude rule stay close together. Above 40 degrees north the two diverge — a third-order polynomial fit of the PVWatts data matches actual optimal tilt better than the linear latitude rule does, and vendor guidance independently puts the real optimum 5–15 degrees flatter than latitude at most locations. Below that band, near the equator, a different rule takes over: minimum tilt is set by rain-cleaning needs, not by irradiance capture. Always confirm with a full site simulation before finalizing.

Q2. How much yield do I actually lose from an off-optimal roof?

Less than most clients expect, within limits. Deviations of up to 15 degrees in tilt and 30 degrees in azimuth cost under 5% of annual yield by one vendor's published table. A full east- or west-facing roof costs more — 20–24% against a south-facing baseline. At temperate latitudes a 45-degree azimuth turn costs 5–8%, more than the same turn would cost closer to the equator. These are vendor-published figures that disagree slightly between sources, so treat them as a planning range, not a guaranteed number for a client contract.

Q3. Why would I ever build flatter than the energy-optimal tilt?

Because on an area-constrained site, total site output matters more than per-panel efficiency. Modelling shows that at 30 degrees of tilt you can fit less than half the ground area in collectors compared with a flatter layout at the same shading limit, and that going from 30 degrees down to 5 degrees lets you install almost twice the PV power on the same plot, losing about 9% of per-panel efficiency but gaining more in total energy. On a flat commercial roof or a density-limited ground plot, this trade usually wins over chasing the theoretical energy-optimal tilt.

Q4. When does a tracker actually pay for itself over fixed tilt?

Trackers make sense where the site doesn't have a specific condition working against them. Fixed-tilt stays the better call in high winds, heavy snow loads and steep topography, and in high-wind zones trackers often stow flat during storms, meaning the site is effectively running as fixed tilt during exactly the conditions a tracker was bought to handle. Trackers also need more land for the same output than fixed tilt does, since fixed tilt can be installed at higher density. If the site has none of those constraints, single-axis tracking is generally the right call — it captures within 1–3% of dual-axis tracking annually, and paying for a second axis rarely makes sense outside of niche cases.

Q5. Is a dual-axis tracker ever worth the extra cost over single-axis?

Rarely, based on the modelled data. Single-axis horizontal tracking comes within 1–3% of dual-axis tracking on incident radiation annually averaged, at most latitudes, while dual-axis systems need more land to keep back rows from shading. Above 75 degrees north or below 60 degrees south, tracking of any kind stops adding much over a well-chosen fixed tilt. Unless a specific site study shows otherwise, single-axis tracking is the point where the incremental yield stops justifying the incremental cost.

Q6. How should row spacing be calculated on a flat roof or ground-mount plot?

Against the worst-case sun angle of the year, not an average day. Row spacing is calculated as module height times the sine of tilt, divided by the tangent of the solar altitude at winter solstice noon — the shorter the shadow the winter sun throws, the tighter you can pack the rows without cross-shading. The same latitude that shifts optimal tilt also shifts this number sharply: a 52-degree-north site with a 30-degree tilt and a 1.7 m module needs about 3.3 m of row spacing, while the same setup at 40 degrees north needs only 1.9 m. Designers use December 21 as the benchmark date because a panel shade-free at noon on that day is shade-free for the rest of the year at solar noon.

Q7. Do I need to worry about magnetic declination when setting azimuth?

Yes, and it is an easy step to skip. A handheld compass reads magnetic south, not true south, and the difference — magnetic declination — can be as much as 1 to 15 degrees or more depending on location. Azimuth for solar design should always be set from true south using GPS coordinates or satellite imagery, not a compass reading, since even a modest declination error compounds with every other tilt and azimuth assumption in the design.

Q8. Should I trust the quick tilt/azimuth optimisation graph in my design software for a client proposal?

No — use it to sanity-check a design direction, then confirm with the full simulation before it goes in front of a client. Documentations warn that its quick orientation estimate is calculated from monthly weather values and can differ from the GlobInc figure the full simulation produces. The same caution applies to any quick-mode tilt/azimuth tool: it is fast enough to steer a design conversation, not precise enough to be the number a proposal is built on.

Sources

• Jacobson and Jadhav, "World estimates of PV optimal tilt angles and ratios of sunlight incident upon tilted and tracked PV panels relative to horizontal panels," Solar Energy (2018), Stanford
• Solar Power World — "Fixed-tilt racking still has a place in utility-scale solar," April 2025
• OSTI — NREL/TP-6A20-64102, System Advisor Model Photovoltaic Performance Model Technical Reference
• Sunlith Energy — Solar Panel Tilt Angle by Location