📦 Contents

Chapter 1: Understanding Desertification and the Land Degradation Neutrality Framework

Desertification is one of the few environmental problems whose name suggests the wrong picture. It does not mean sand dunes advancing across a map, and it is not a synonym for drought. It is a slow loss of the land's capacity to support vegetation, soil life, water retention and the people who depend on them, taking place in regions that are dry to begin with. Because that loss is gradual, patchy and entangled with ordinary year-to-year swings in rainfall, it is hard to see without careful, repeated and comparable measurement. This chapter explains what is being measured and why, and it prepares the ground for the rest of the book, which shows how the WIA-ENE-057 standard turns those measurements into interoperable records, APIs, protocols and verifiable evidence.

The chapter has no code. Its goal is conceptual: by the end, a GIS engineer, a backend developer or a restoration-project data manager should be able to explain the phenomenon to a colleague, name the international framework that governs how countries report on it, and say which part of the WIA data model carries each kind of evidence. Readers who prefer to see data first can open the first tab of the companion simulator (“1. Data Format”, linked below this chapter), which renders example vegetation, soil and rainfall records for the running example location SAHEL-REGION-001.

1.1 What Desertification Is — and What It Is Not

The convention definition

The reference definition comes from Article 1 of the United Nations Convention to Combat Desertification (UNCCD): desertification is land degradation in arid, semi-arid and dry sub-humid areas resulting from various factors, including climatic variations and human activities. Three elements of that sentence matter for anyone designing a data system.

• It is a kind of land degradation. The object being tracked is a reduction in the biological or economic productivity of land — rangeland that carries less forage, cropland that yields less per unit of rain, woodland that has thinned to scattered shrubs. Degradation is a change relative to a reference state, so the data model must always be able to express “compared with what”.
• It is geographically bounded. The same loss of productivity in a humid tropical forest is land degradation, but it is not desertification in the convention's sense. A data platform must therefore know which climatic zone a monitored location belongs to, which in turn depends on long-term climate data rather than on a single season.
• It has mixed causes. Climate variability and human activity act together. A record that attributes a decline to “drought” alone, or to “overgrazing” alone, is usually an oversimplification; the standard therefore keeps climatic evidence and human-pressure evidence in separate, joinable records instead of collapsing them into one cause code.

Drylands and the aridity index

The climatic scope is set by the aridity index, defined as the ratio of long-term mean annual precipitation (P) to long-term mean annual potential evapotranspiration (PET):

AI = P / PET        (both in mm per year, long-term means)

PET is the amount of water that a well-watered reference surface would evaporate and transpire under the prevailing radiation, temperature, humidity and wind. It is not measured directly but computed, most commonly with the FAO-56 Penman–Monteith method or, where only temperature is available, with simpler empirical formulas. A location whose rainfall is a small fraction of what the atmosphere could take up is water-limited for most of the year, and that is what makes its vegetation and soils fragile. The widely used classification, introduced in the UNEP World Atlas of Desertification, is summarised below.

For implementers, two consequences follow. First, the aridity class is an attribute of the location, derived from climate normals, not of any individual observation; it belongs with the location's registration metadata rather than being recomputed from each season's rainfall. Second, because the class boundaries are thresholds on a continuous quantity, locations near a boundary can shift class when normals are updated or when a different PET method is used. A platform should therefore store the method and the reference period alongside the class so that a later reader can reproduce it.

Not a natural desert, not a drought

Two confusions recur in project documents and press material, and both lead to bad data design.

Desertification is not the existence of deserts. A hyper-arid desert with sparse vegetation is in its natural state; low vegetation index values there are not evidence of degradation. If a platform raised alerts purely on absolute thresholds — “NDVI below some value means desertified” — it would flag every healthy desert and miss a semi-arid savanna that has lost a third of its productivity but is still greener than the desert. The WIA model avoids this by expressing condition as trends and comparisons (trendAnalysis in vegetation records, historicalComparison in rainfall records) rather than as absolute labels alone.

Desertification is not drought. Drought is a temporary shortfall of water relative to the normal for that place; when rains return, vegetation on healthy land recovers. Desertification is what remains after the rains return: soils that have lost structure and organic matter, seed banks that have been depleted, gullies that now shed water instead of storing it. Drought can trigger or accelerate degradation, and degraded land suffers more from the next drought, but the two must be measured differently. That is why the standard keeps a dedicated rainfall record with a drought index alongside, not inside, the vegetation record: an analyst needs both to tell a dry year from a damaged landscape.

1.2 Drivers and Feedback Loops

Climatic drivers

In drylands, rainfall is not only low but also highly variable between years and concentrated in a short wet season, often in intense convective storms. Several climatic factors raise the risk of degradation:

• Rainfall deficits, especially multi-year runs of below-normal seasons, which reduce plant cover and allow soils to dry and loosen.
• Heat, which raises evaporative demand and so lowers the effective water available to plants even when rainfall is unchanged.
• Intense rainfall events on bare ground, which cause splash erosion, surface sealing and runoff rather than infiltration.
• Strong dry-season winds, which remove fine, nutrient-rich particles from exposed soil.

These factors explain why the standard's rainfall record captures not just the seasonal total but also its distribution and intensity, and why the soil record includes wind-related erosion factors. A total that looks normal can hide a season in which most rain fell in a few storms on bare soil.

Human drivers

Human pressure acts on the same land, usually through livelihoods that are rational at the household level but damaging when they exceed what the land can sustain:

• Overgrazing: stocking rates above the carrying capacity of rangeland, particularly when herds are held near water points during the dry season.
• Cropland expansion into marginal land and shortening of fallow periods, leaving soils uncovered for longer.
• Fuelwood and charcoal collection, which removes woody cover that anchors soil and moderates microclimate.
• Inappropriate tillage and irrigation: repeated ploughing that breaks soil structure, and irrigation without drainage that leads to salinization.

Many of these drivers are observable, at least indirectly: stocking rate from livestock counts and grazing area, cropland extent from land-cover mapping, fuelwood pressure from surveys. The WIA land-use record stores them explicitly so that an assessment can separate a climate-driven decline from one where pressure has risen.

The self-reinforcing loop

What makes desertification persistent is that its physical processes reinforce each other. Once vegetation cover falls below a certain level, the land starts to lose the very properties that would allow vegetation to return. The loop is shown in Figure 1-1.

   Climatic stress (rain deficit, heat)      Human pressure (grazing, clearing, tillage)
                   │                                        │
                   └──────────────────┬─────────────────────┘
                                      ↓
                        ┌──────────────────────────────┐
             ┌────────→ │ 1. Loss of vegetation cover  │
             │          └──────────────────────────────┘
             │                        ↓
             │          ┌──────────────────────────────┐
             │          │ 2. Bare soil exposed to      │
             │          │    raindrop impact and wind  │
             │          └──────────────────────────────┘
             │                        ↓
             │          ┌──────────────────────────────┐
             │          │ 3. Surface crusting, loss of │
             │          │    organic matter, wind and  │
             │          │    water erosion             │
             │          └──────────────────────────────┘
             │                        ↓
             │          ┌──────────────────────────────┐
             │          │ 4. Lower infiltration, more  │
             │          │    runoff, less stored water │
             │          └──────────────────────────────┘
             │                        ↓
             │          ┌──────────────────────────────┐
             └───────── │ 5. Less water and nutrients  │
                        │    for plant regrowth        │
                        └──────────────────────────────┘

Two properties of this loop shape how monitoring should work. The first is lag: soil properties change more slowly than vegetation indices, so a greening response after one good wet season may sit on top of soils that are still deteriorating. The second is threshold behaviour: land can absorb pressure for years with little visible change and then shift relatively quickly to a degraded state from which recovery requires active intervention. Both argue for keeping long, consistent time series and for combining several indicators instead of relying on any single one. They also explain why the WIA model carries a confidence value alongside trend statuses: a short or noisy series should not be reported with the same assurance as a long, stable one.

The loop can also be run in reverse, which is the basis of restoration. Measures that capture water and hold soil in place — micro-catchments, stone lines, windbreaks — raise infiltration at stage 4, which supports regrowth at stage 5, which in turn protects the soil at stage 2. Monitoring that is designed well can therefore show not only decline but also the early signs of recovery.

1.3 The Four Observable Pillars the Standard Measures

Desertification cannot be measured directly; it is inferred from a set of observable variables. WIA-ENE-057 groups those variables into four pillars — vegetation, soil, climate and human pressure — and assigns each pillar its own record type in PHASE 1. Two further record types describe the human response (restoration projects) and the synthesis (monitoring reports). All records share a locationId, which is the join key that lets an analyst line up, for one place and one period, what the vegetation did, what the soil did, how much it rained and how intensively the land was used.

The separation is deliberate. A single “desertification score” stored without its inputs cannot be audited or recomputed when a method improves. By keeping each pillar in its own record, with its own source, resolution and quality metadata, the standard lets a reviewer trace any composite figure back to the observations behind it. It also lets different organisations contribute different pillars: a national meteorological service can supply rainfall records, a forestry agency vegetation records, and an NGO restoration records, all joined by the same location identifier.

Each pillar also has its own natural cadence. Satellite vegetation composites typically arrive every few weeks, soil-moisture probes can report several times a day, rainfall is summarised by season and year, and land use changes on a scale of years. Chapter 2 explains the science behind each indicator, and Chapter 3 specifies the fields in detail. Throughout, the numbers quoted from the specification's examples — such as an NDVI value of 0.35 for SAHEL-REGION-001 — are illustrative example values, not observations of any real place.

1.4 UNCCD, SDG 15.3.1 and Land Degradation Neutrality

From convention to global target

The UNCCD, adopted in 1994, is the international treaty that addresses desertification, land degradation and drought. Its Parties report periodically on the state of their land and on the implementation of the convention. In 2015 two developments gave that reporting a common quantitative target. The 2030 Agenda for Sustainable Development included SDG target 15.3: by 2030, combat desertification, restore degraded land and soil, and strive to achieve a land degradation-neutral world. In the same year the UNCCD Conference of the Parties adopted the concept of Land Degradation Neutrality (LDN), defined as a state whereby the amount and quality of land resources necessary to support ecosystem functions and services and enhance food security remain stable or increase within specified temporal and spatial scales and ecosystems.

The indicator attached to the target, SDG 15.3.1, is the proportion of land that is degraded over total land area. The UNCCD is the custodian agency for this indicator, and countries report it through the convention's national reporting process.

The three sub-indicators

Rather than inventing a single composite, the LDN framework assesses degradation through three sub-indicators that together cover the physical, biological and carbon dimensions of land:

1. Land cover — change between land-cover classes (for example, from tree-covered area or grassland to cropland or bare land). Certain transitions are classed as degradation, others as improvement.
2. Land productivity — usually expressed as land productivity dynamics, a classification of long-term trends in net primary productivity, which in practice is estimated from time series of satellite vegetation indices.
3. Carbon stocks above and below ground — in current practice represented by soil organic carbon (SOC) stocks, which change slowly and reflect the long-term health of the soil.

Baseline and “one out, all out”

LDN is a comparison against a baseline: each country establishes the state of its land for an agreed reference period and then assesses later periods against it. Neutrality does not require that no hectare ever degrades; it requires that losses are balanced by gains, ideally within the same land type, so that the total amount and quality of land is at least maintained.

To decide whether a particular land unit is degraded, the framework applies the “one out, all out” principle: if any one of the three sub-indicators shows significant negative change, the unit is counted as degraded, even if the other two are stable or improving. This conservative rule prevents, for example, a rise in vegetation productivity from masking a loss of soil carbon. It has a direct consequence for data design: a platform that supports LDN reporting must keep the three sub-indicators separately evaluable per land unit, with their own baselines and significance tests, and must not average them into one score before the rule is applied.

Where WIA-ENE-057 fits

WIA-ENE-057 does not replace the convention's reporting process or its methodological guidance. Its role is narrower and practical: to make the underlying evidence interoperable so that it can be assembled into LDN-style reports with less manual work and with a clear trail back to the observations. Concretely:

• Vegetation records supply the productivity time series; land-use records supply the land-cover classes and their areas; soil records and carbon-related biomass fields supply the inputs for carbon-stock estimates.
• The PHASE-2 API defines POST /api/v1/reports/unccd-ldn, which assembles baseline and current values for land cover, productivity and carbon stocks for a set of locations into a downloadable report (Chapter 4).
• The composite monitoringReport carries a certifications object with flags such as ldnCompliant and unccdReported.

These flags need careful interpretation. ldnCompliant is a statement made by the implementation about its own data package — that the required sub-indicators are present and computed as documented — not a finding by the UNCCD about a country's or project's neutrality status. Likewise, any report generated by a WIA endpoint is an input to national reporting, which remains the responsibility of the reporting Party through the convention's own channels. Chapter 6 returns to this boundary when it describes the integration layer in detail.

1.5 Restoration Response Options

Field techniques

A monitoring standard is only useful if it also records what people do about the problem, so that the effect of interventions can be evaluated. Dryland restoration draws on a well-established toolbox. The techniques below are those most often encountered in Sahelian and similar contexts; several of them appear by name in the specification's example restorationProject.

• Half-moon (demi-lune) micro-catchments: semicircular earth bunds dug on gentle slopes with the open side facing upslope, which trap runoff and sediment so that a planted tree or crop receives far more water than rainfall alone would provide. Recorded as a water-management technique ("half-moon-technique").
• Stone bunds: lines of stones laid along contours that slow runoff, trap sediment and seeds, and gradually build up fertile strips ("stone-bunds").
• Planting pits (known in parts of West Africa as zaï): small pits filled with organic matter that concentrate water and nutrients around seeds.
• Rainwater harvesting at household or landscape scale, from roof and surface catchment to small reservoirs ("rainwater-harvesting").
• Windbreaks and green belts: rows of trees and shrubs that reduce wind speed at the soil surface, cutting wind erosion and evaporation on the protected side.
• Assisted natural regeneration, including farmer-managed natural regeneration, in which existing rootstocks and stumps are protected and pruned so that trees regrow without replanting.
• Reforestation with drought-tolerant species, preferably native. The specification's example names Acacia senegal, the source of gum arabic, and Balanites aegyptiaca, the desert date; both are native to the Sahel and valued for products as well as for cover.
• Grazing management: rotational grazing, seasonal closures and adjusted stocking rates, which address a driver rather than a symptom.

For data purposes, the important point is that each of these interventions has a measurable footprint — area treated, structures built, trees planted and surviving, change in soil moisture or vegetation index in treated versus untreated plots. The restorationProject record is structured to capture those footprints (species with quantities and survival rates, soil-conservation measures, water-management techniques, milestones and outcomes), so that a claim such as “the site is recovering” can be checked against the same vegetation and soil records used to detect decline.

Programme context

Many restoration projects in drylands sit within larger initiatives. The Great Green Wall, led by the African Union with countries across the Sahel, aims to restore degraded land in a broad band south of the Sahara; over time it has evolved from the image of a continuous tree line to a mosaic of land-management practices. The Bonn Challenge is a global call for forest landscape restoration pledges, and AFR100 is an African country-led initiative contributing to it. These programmes are relevant to the standard because projects often report to several of them at once, which multiplies the risk of inconsistent or double-counted figures. This book deliberately does not quote programme-wide hectare or tree targets; readers should consult each programme's own publications, and Chapter 6 treats such programmes purely as integration partners in a data sense.

Korean context

Korea is a Party to the UNCCD and has a particular connection to the convention. The tenth session of the UNCCD Conference of the Parties (COP10) was held in Changwon in 2011, and the host country launched the Changwon Initiative there to support implementation of the convention, including work on scientific approaches to land degradation and partnerships for sustainable land management. Korea's own history of large-scale reforestation after the mid-twentieth century is often cited as an example of how degraded hillsides can be restored through sustained policy, and the Korea Forest Service has pursued dryland afforestation cooperation with partner countries, with Mongolia as a prominent example. For Korean agencies, companies and NGOs working on such cooperation, a common data standard offers a way to report results in a form that partner governments and international reviewers can verify, rather than in project-specific formats that are hard to compare.

1.6 Why a Data Standard Is Needed

The science and the policy framework described so far are mature. What remains difficult is the plumbing: getting observations from many sources into a form in which they can be combined, compared over time and trusted by someone who did not collect them. Four problems recur.

Fragmented satellite products. Vegetation indices are available from several missions at different spatial resolutions, revisit intervals, compositing periods and quality-flag conventions. A time series that silently mixes a coarse 16-day composite with a finer single-date scene can show a spurious trend. Unless each value carries its source, resolution and quality information, downstream users cannot tell.

Incompatible field forms. Soil samples, moisture readings, livestock counts and tree-survival surveys are often collected on project-specific forms, with local units, inconsistent location references and no shared vocabulary. Merging data from two projects in the same district can take longer than collecting it.

Unverifiable restoration claims. Numbers such as trees planted or hectares restored are frequently reported without the survival rate, the location geometry or the monitoring data that would allow anyone to check them. When the same trees are reported to several programmes, totals inflate. This undermines trust in restoration as a whole.

Heavy reporting burdens. National focal points, donors and, increasingly, carbon-market participants each ask for similar information in different templates. Without a shared structure, every report is assembled by hand, and errors introduced in one become facts in the next.

WIA-ENE-057 addresses these problems at four layers, each of which is the subject of later chapters:

• Common JSON records with explicit identifiers, timestamps, coordinates (WGS84), units, sources and quality metadata, and validation rules such as rejecting an NDVI value outside −1 to 1.
• A REST API and streaming interfaces for registering locations, submitting observations, computing risk assessments, planning restoration and generating reports.
• Protocol rules for real-time sensor data, satellite ingest, OGC map services, webhooks with signatures, bulk upload and security.
• Integration patterns and signed conformance evidence, so that an implementation can show — not just claim — that it follows the standard.

It is equally important to state what a data standard cannot do. It cannot make a poor measurement accurate, choose the right restoration technique for a given slope, or settle policy disputes about land tenure and grazing rights. It can make the evidence consistent, traceable and reusable, which is a precondition for all of those discussions to be grounded in the same facts.

1.7 The Four PHASEs and How This Book Follows Them

WIA-ENE-057 is published as four PHASE documents, each building on the one below it. PHASE 1 defines what the data looks like; PHASE 2 defines how it is exchanged through an API, including a separate API-INTERFACE document that specifies the discovery document, the signed manifest and the conformance tiers; PHASE 3 defines the protocols over which data moves; and PHASE 4 describes how an implementation connects to external systems such as Earth-observation archives, GIS platforms, restoration programmes and reporting workflows. Figure 1-2 shows the stack and where each chapter of this book sits.

┌──────────────────────────────────────────────────────────────┐
│ PHASE 4  Integration                                          │  Ch. 6  Global ecosystem
│   LDN reporting · EO pipelines · GIS · field apps · carbon    │  Ch. 7  Architecture & deployment
├──────────────────────────────────────────────────────────────┤
│ PHASE 3  Protocol                                             │  Ch. 5  Protocols & real-time flow
│   WebSocket · MQTT · satellite ingest · OGC · webhooks · TLS  │
├──────────────────────────────────────────────────────────────┤
│ PHASE 2  API  +  API-INTERFACE                                │  Ch. 4  REST API & risk assessment
│   /monitoring · /assessment · /restoration · /reports         │
│   discovery document · signed manifest · conformance tiers    │  Ch. 8  Conformance & certification
├──────────────────────────────────────────────────────────────┤
│ PHASE 1  Data format                                          │  Ch. 3  Data format
│   vegetationData · soilData · erosionData · rainfallData      │
│   landUseData · restorationProject · monitoringReport         │
├──────────────────────────────────────────────────────────────┤
│ Domain foundations                                            │  Ch. 1  Desertification & LDN
│   drylands · drivers · indicators · LDN framework             │  Ch. 2  Remote-sensing & field science
└──────────────────────────────────────────────────────────────┘

The book follows the stack from the bottom up:

• Chapter 2 explains the science behind the indicators: how vegetation indices are computed, what the main satellite products offer, how soil, erosion and rainfall are measured, and how to detect trends without mistaking a wet year for recovery.
• Chapter 3 specifies the PHASE-1 records field by field, with validation rules, a JSON Schema fragment and advice on units and schema evolution.
• Chapter 4 covers the PHASE-2 REST API, including the location-first workflow, the weighted desertification risk score, restoration plans, alert subscriptions, LDN-style reports, error handling and the discovery document.
• Chapter 5 turns to PHASE 3: WebSocket streaming, MQTT sensor topics, satellite raster ingest, OGC services and their axis-order pitfalls, webhook signature verification, bulk upload and protocol-level security.
• Chapter 6 describes PHASE-4 integration with reporting frameworks, Earth-observation pipelines, agricultural and soil partners, restoration networks, GIS and field apps, and carbon accounting — with care about what such integrations can and cannot claim.
• Chapter 7 assembles these pieces into a reference architecture and follows one illustrative season at SAHEL-REGION-001 from sensor to annual report.
• Chapter 8 closes with conformance testing: the Surface, Verified and Anchored tiers, test vectors, evidence packages, the command-line tool, versioning and governance, and the directions in which the field is moving.

A single running example ties the chapters together. The location SAHEL-REGION-001 (latitude 14.5, longitude −4.0, 10,000 hectares) appears in every PHASE document with a consistent family of identifiers — VEG-20251225-SAHEL-001 for a vegetation record, SOIL-20251225-SAHEL-001 for a soil record, RAIN-2025-SAHEL-001 for a rainfall season, REST-2025-SAHEL-001 for a restoration project and REPORT-2025-Q4-SAHEL-001 for a quarterly report. The location and all its values are fictional teaching material. Where the specification's example values are inconsistent with one another, later chapters point this out rather than repeating them as if they were measurements.

Chapter Summary

• Desertification is land degradation in arid, semi-arid and dry sub-humid areas (aridity index 0.05 to below 0.65), caused by climatic variation and human activity together. It is neither the existence of natural deserts nor a synonym for drought.
• Degradation persists through a feedback loop — vegetation loss, bare soil, crusting and erosion, lower infiltration, less regrowth — which also shows where restoration can intervene.
• WIA-ENE-057 measures four pillars (vegetation, soil, climate, human pressure) in separate records joined by locationId, plus records for restoration responses and composite monitoring reports.
• SDG indicator 15.3.1 and the LDN framework assess degradation through land cover, land productivity and carbon stocks (SOC), against a baseline, using the “one out, all out” rule. WIA-ENE-057 packages evidence for this framework but does not replace national reporting.
• A data standard is needed because satellite products, field forms and restoration claims are fragmented and hard to verify; the four PHASEs address data format, API, protocol and integration, and the book follows them in order.