How we screen a renewable energy site before it ever reaches feasibility

Developers rarely lose money on sites with bad resource. They lose money on sites with good resource and everything else wrong — an interconnection queue that adds three years, a land parcel split across incompatible tenure types, or a protected-area buffer nobody checked before the option agreement was signed. Site suitability screening exists to catch that class of problem early, cheaply, and before it becomes a feasibility-study write-off.

Resource is the easy part

Solar irradiance and wind speed data are, at this point, commoditised. Reanalysis datasets, satellite-derived irradiance products, and mesoscale wind models are publicly available or cheaply licensed, and most competent teams can produce a defensible resource estimate for almost anywhere on the planet in an afternoon. That is exactly why resource is not where screening should spend its time — it is necessary, but it separates almost nothing.

What actually separates a bankable site from a dead one sits underneath the resource layer: constraints that are binary rather than gradual. A parcel is either connectable within an acceptable cost envelope or it isn't. Land is either clean or it carries a dispute. A route to grid either exists or it has to be built. Screening has to test the binary constraints before it spends effort refining the continuous ones.

The constraint stack, in the order we actually check it

• Land and tenure. Ownership records, competing claims, protected and reserved categories, and whether the parcel shape and topology can physically host the layout being proposed.
• Grid proximity and capacity. Distance to the nearest substation or line of adequate voltage, and — critically — whether that substation has published or inferable headroom, not just proximity on a map.
• Environmental and social constraints. Protected areas, biodiversity corridors, watercourse buffers, and proximity to settlements that trigger a higher class of environmental or social impact assessment.
• Physical and engineering constraints. Slope, soil bearing capacity for foundations, flood risk, and access-road feasibility for construction traffic.
• Regulatory and permitting pathway. Which consenting regime applies, how long it typically takes in that specific jurisdiction and technology class, and whether any moratoria or zoning restrictions currently apply.

Each layer is checked as a filter, not a score, in early screening. A site that fails land tenure does not get partial credit for excellent wind. It gets removed, because no amount of resource quality fixes an unresolvable land dispute or a substation with no available capacity.

Why the order of filters matters

Running the cheapest, most binary filters first is a cost discipline, not a technical preference. Land tenure and grid-capacity checks are comparatively fast to run against public and semi-public records. Environmental and social screening, by contrast, often requires desk-based GIS overlay work against protected-area and biodiversity datasets that is more time-intensive to do properly. Running the expensive filters against a universe that has already been thinned by the cheap ones is simply a better use of analyst hours — and it is the difference between screening 40 candidates in two weeks and screening 40 candidates in two months.

The other reason for sequencing this way is that grid and land constraints are the ones most likely to be irreversible on a normal project timeline. A permitting delay can sometimes be worked around with a revised timeline. A substation with no capacity headroom, or a land parcel with an unresolved title dispute, usually cannot be worked around at all — so those checks belong as early gates rather than late-stage due diligence items.

What this looks like in practice

On a typical site-prospecting mandate, we start from a resource-screened candidate list — sometimes generated by the client, sometimes by us — and run it through the constraint stack using a mix of GIS overlay analysis, public land registries, utility interconnection queue data where published, and jurisdiction-specific regulatory research. The output is not a single suitability score. It is a shortlist with a documented reason for every site that was excluded, because the client's internal stakeholders and, eventually, their lenders will ask why a site was dropped — and "the model gave it a low score" is not an answer that survives that conversation.