Terrain-Aware Backtracking: How to Optimise Tracker Layouts on Undulating Terrain

BTR Enerji Engineering Team · Published: 29 September 2026 · 4 min read

Most single-axis tracker plants are designed as if the land were flat, but sites rarely are. A few tens of centimetres of height difference between rows can turn into morning and evening shading, and that shading into a disproportionate electrical loss. This article explains the two ways slope affects a tracker and how terrain-aware backtracking reduces the loss.

Two components of slope: along the axis and across it

The tracker axis runs north-south, so terrain slope splits into two components:

• Along-axis slope (north-south): the table follows it and the axis becomes tilted. In the northern hemisphere an axis falling towards the south slightly increases annual yield; one falling towards the north reduces it. Tracker products typically accept around 10–15% axis slope; steeper areas need grading.
• Cross-axis slope (east-west): neighbouring rows sit at different heights. A row that is lower on the sun-facing side falls into its neighbour's shadow earlier than it would on flat ground.

Why standard backtracking is not enough

Backtracking rotates the tables away from the sun at low sun angles so that rows do not shade each other. The standard algorithm assumes all rows lie in one plane at equal spacing. With cross-axis slope that assumption fails: shading appears at angles the algorithm considers shade-free, or tables back-track more than needed and lose irradiance. NREL's slope-aware backtracking work (Anderson and Mikofski, 2020) corrects the geometry with the cross-axis slope angle, and the open-source pvlib library implements it.

Row-by-row or site-wide?

When slope differs for every pair of rows, a single correction does not fit all. There are two approaches:

• Site-wide: one average slope is applied to the whole site. Simple, but residual shading remains wherever a row pair deviates from the average.
• Row-by-row: each table rotates according to the height difference to its own neighbour. Residual shading is much lower; the control system must support per-row parameters.

Our simulator computes both modes, so you can see what row-level control is worth on your site before the investment decision.

Small shade, large loss: the electrical effect

On a 1P (one module in portrait) tracker, row shading starts at the lower edge of the module and moves up along it. Even a thin strip of shade can affect several cell strings; once bypass diodes activate, the loss is far larger than the shaded area fraction. Treating shaded area as a pure irradiance loss is therefore misleading. Our model evaluates the electrical effect of residual shade at cell-height scale: millimetre-level shade is not counted as a full loss, and shade covering a whole cell row is not underestimated.

Which elevation data is good enough?

• Global elevation models (EU-DEM 25 m, SRTM 30 m): good for the overall axis slope and large-scale terrain. They only roughly capture height differences across a 5–7 m row pitch, so very small differences are treated as noise and filtered out.
• Topographic survey (drone photogrammetry, total station, GNSS): the right input for row-level shading and pile lengths. You can upload it to the simulator as a DEM, GeoTIFF or DXF with elevations.

Practical design recommendations

• Identify areas beyond the axis-slope tolerance early and include grading cost in the feasibility study.
• Where cross-axis slope is high, a slightly lower ground coverage ratio combined with row-by-row backtracking gives the best result.
• Set pile lengths from survey data; clearance under the table matters for both bifacial gain and maintenance.
• Ask your tracker supplier whether the controller supports per-row slope parameters.

Slope limits and earthwork: Solar tracker slope tolerance.

Frequently asked questions

How much does terrain-aware backtracking gain?

It depends on the site. On flat land the difference is negligible; on sites with variable cross-axis slope it clearly reduces morning and evening row shading. The simulator compares both modes to show the difference for your land.

What is the maximum slope for a single-axis tracker?

Along the axis (north-south) the typical limit is around 10–15% and varies by product. East-west slope does not prevent installation but increases row-to-row shading.

Is satellite elevation data sufficient?

It is sufficient for the overall slope at pre-feasibility stage. Row-level shading, pile lengths and earthwork need a site survey.

The ratios on this page are typical industry ranges and are not binding; exact values are determined by a site-specific analysis.