Kingdom of Saudi Arabia · National Land Evaluation · Application 2

Database Application

The data layer of the national agricultural land evaluation. Six modules covering the logical design, where every value comes from, how satellite observations become stored columns, what the ministries must supply, how far each result can be relied on, and what is missing. Specification only — no physical implementation exists.

Modules7

Tables39

Columns456

Elements66

GridHexagonal · 1 ha

Modules

Built

Schema

The logical design

Eight schemas, thirty-nine tables, four hundred and fifty-six columns, with a sample value on every column. Design decisions, the lookup register, class ranges and a nine-hop lineage trace.

What the study cannot obtain elsewhere

Ten datasets across nine authorities, itemised as a formal request with the attributes each one must carry, what it blocks, and the critical path through them.

Where each element can come from

All 64 elements mapped to candidate datasets, ranked primary, alternative and fallback, with the resolution score each source fixes before any processing happens.

Sensor to stored column

Fourteen derivation chains with every processing step explained, four sensors with their bands and indices, and the cloud, compositing and resampling rules each chain depends on.

How far a result can be relied on

The four-component model, the ceiling each element can reach given the source chosen for it, and what calibration would buy against what field measurement would buy.

What exists, and what is missing

Build status against expected national coverage, the eight quality measures defined before there is anything to measure, and every open gap in one register.

What a hectare actually returns

Six illustrative cells. Select one to see its four class codes, expand any module to the elements and values behind it, and read how the result is interpreted — including what governed it and what is unknown.

Which module answers which question

Where does this value live?Schema → Tables. Every column with its type, key, a sample value and its definition.

What class does this number give?Schema → Class ranges. 54 elements with the table and column each one tests.

Where do I get this data?Data Sources. Primary, alternative and fallback for each of the 64 elements.

How is it processed?Remote Sensing. Fourteen chains, every step explained, with where each is weakest.

Why can we not start?Ministry Data for what must be requested; Data Quality for everything else that is open.

What does a result look like?Sample Output. Six illustrative cells with their class codes, elements and interpretation.

How much can we trust it?Confidence Scoring. The model, the ceiling per element, and what would lift it.

Two ceilings bind every output regardless of build progress. No threshold version exceeds source_validated, capping confidence component 4 at 0.80 everywhere; and no element is field-measured anywhere, so component 1 cannot reach 1.00. No cell can currently publish a confidence above roughly 0.92. This is a property of the study as scoped — satellite and desk-based, with no field survey — not a defect in the design.