Assisted Natural Regeneration & Native Enrichment in Kumaon: A Site-Assessment Framework

The Western Himalaya — encompassing the mid-montane zone (1,000–2,500 m) and the subalpine belt (2,500–3,500 m) — lies within the Himalaya Biodiversity Hotspot, one of the world's 36. Across the hotspot, CEPF records about 10,000 plant species, roughly 30% of them endemic, and its forests and soils hold large carbon stocks.

The two survival figures are not contradictory: early counts on new plantings and later counts on older ones measure different things. The Green Credit Programme awards credits on trees that are still alive after at least five years, so the later figure is the relevant one.

The 1,001–2,000 m altitudinal band is a high-priority intervention zone: exposed to encroachment and fire, yet often retaining ecological memory — seed banks, coppice regrowth, mycorrhizal fungi — that makes Assisted Natural Regeneration a sensible first response over clean-slate plantation where that memory is present.

ANR uses residual ecological potential — surviving root stocks, seed banks, nurse trees — in degraded but not ecologically dead landscapes. Where residual rootstock and seed sources survive, ANR is often substantially cheaper than planting (a modelled 77% saving in Brazil's Atlantic Forest, Crouzeilles et al., 2020), though the saving is site-dependent (Busch et al., 2024). ANR is not passive: it needs a site diagnosis followed by deliberate facilitating work. Whether it earns Green Credits is a separate question (see the eligibility note below).

Stage 1 — Ecological Memory Inventory (EMI)

Thresholds are provisional design values to be calibrated locally. Plots should be laid out as a stated sample design, so counts carry an interval.

Memory class and site type are separate choices. The EMI score (high, medium or low memory) decides how much planting is needed. The site type (A, B or C below) decides which operations apply.

Protocol A — Invasive-Dominated Degraded Forest

Where Lantana camara or Ageratina adenophora (formerly Eupatorium adenophorum) cover exceeds 60%, the invasive-first sequence applies before any ANR or planting.

Protocol B — Fire-Maintained Chir Pine Grassland

Pinus roxburghii-dominated landscapes are fire-climax types with suppressed broadleaf regeneration. ANR here makes fire exclusion the primary driver, not plantation.

Protocol C — Degraded Agroforestry / Van Panchayat Margins

Partially vegetated, human-influenced sites — field boundaries, terrace margins, Van Panchayat edges. Hybrid ANR-enrichment approach with soil work before planting. The rows below are this framework's working prescriptions. They are not Green Credit or State requirements, and several have little field evidence (see the evidence box that follows).

A single species prescription applied across a hillslope without micro-climate mapping risks high mortality in unsuitable microsites; we have no Kumaon measurement of how high, so treat this as a design assumption. All sites > 2 Ha should have the following parameters mapped before species assignment (moisture and hydro-physical mapping methods are discussed in Precision Sonar Lake Monitoring).

Species–Microsite Matching Matrix (1,000–2,500 m)

A starting matrix, not a species list. On Green Credit parcels the species list is the one specified by the State/District Nodal Officer. Check each species against that list and against elevation limits; stock should be raised from seed collected in the same elevation band and aspect.

Cedrus deodara has a limited natural range in Kumaon and is mostly planted; Juglans regia is largely cultivated. Use them only where the nodal officer's list includes them. Chir pine is a fire risk, not a restoration species.

Satellite check of the aspect rule (screening only, unassessed): over the Kosi-Katarmal study box (Almora), mean Sentinel-2 post-monsoon NDVI on vegetated slopes of 5° or more was 0.64–0.66 on south-west-facing slopes and 0.70–0.74 on north- and north-east-facing slopes, across four years (2019–2025). The sampling error within a year is about ±0.01, but the comparison is not assessed, and the difference is not evidence that S/SW faces are drier. We apply no further topographic correction: the Sentinel-2 surface reflectance appears already terrain-flattened by ESA's processor, and in our tests an extra per-scene SCS+C correction moved north-slope NDVI further from what GEDI lidar canopy height and NISAR L-band radar, which do not depend on sunlight, indicate (internal analysis, 2026, not peer reviewed). Uncorrected NDVI also differs systematically by aspect: at matched canopy height, north-facing slopes read higher, by roughly 0.04–0.22 NDVI across our five study boxes, and a simple "persistent greenness" rule flags tall forest far more often on north slopes than on south slopes. The cause is not settled: aspect co-varies with forest type and moisture (closed-canopy oak and mixed broadleaf on north slopes, open-canopy chir pine over seasonal grass on south slopes), and a high-sun check in one box suggests direct illumination is not the main driver; forest type, moisture and phenology have not been separated. North-slope NDVI change is therefore treated as unassessed. No field reference sample exists. The figures are consistent with, but do not test, the matrix above. See the forestwatch research preview.

A layered native structure is an ecological aim, supported by the general finding that agroforestry can raise soil carbon, with large uncertainty (Lorenz & Lal, 2014). On Green Credit parcels the rules narrow it: only native species are allowed, credits are counted on new trees older than five years, canopy density is measured by a method the Administrator has not published, and produce stays with the State Forest Department. Set tree density for the 40% canopy test and the tree count first. The shrub and herb layers add biodiversity, not credits, and probably not canopy density.

JICA-assisted Uttarakhand Forest Resource Management Project (UFRMP)

UFRMP (JICA loan signed April 2014; eco-restoration plus livelihoods) is the closest local precedent for restoration on Van Panchayat land. The figures we have for it come from a project brief and are unverified against UFRMP documents: a degraded-forest rehabilitation target of 38,000 Ha split across three Eco-Restoration Models (ERM 1: 20,000 Ha; ERM 2: 9,500 Ha; ERM 3: 8,500 Ha), and 481 selected Van Panchayats in Kumaon, for example Kosi (12 VPs), Gananath (16) and Jageshwar (16) in Almora Civil & Soyam Division, and Mukteshwar (23), Okhalkanda (25) and Ramgarh (31) in Nainital Soil Conservation Division. Check the project's published area-and-targets page before relying on them, and follow the project's ERM guidelines for site prescriptions.

The Wadi Model — Households' Own Land Only

Developed by the BAIF Development Research Foundation in Gujarat and scaled up by NABARD (from 1995 under the KfW-NABARD Adivasi Development Programme in Gujarat), the Wadi model deploys intensively managed multi-species orchard and homestead plots of roughly 0.4–1.0 Ha per household, mainly with tribal and hill households. It applies to land a household controls. It is not a design for State forest land under the Green Credit Programme. We have not found measured Wadi yields or incomes for the mid-Himalaya, so we give none; earlier yield and income tables in this note were illustrative and have been removed.

What Green Credit Parcels Can and Cannot Pay

Per the 1 September 2025 Office Memorandum and its MoU and DPR annexures (our reading; confirm with the nodal officer): the MoU creates no usufruct, ownership or commercial interest for the applicant, produce continues to vest in the State Forest Department, and the applicant may not derive commercial benefit from the parcel. The DPR does ask for NTFP-yielding and pollinator species and records community dependency (fuelwood, grazing, NTFP). Forest law may also treat horticulture and cultivation of medicinal plants on forest land as a non-forest purpose; take legal advice on the current Van (Sanrakshan Evam Samvardhan) Adhiniyam. In practice:

• Aftercare income comes from wages, from produce that existing right-holders may already lawfully take under State or Van Panchayat rules, or from activity off the parcel such as bee-keeping, timur or aromatics on households' own land.
• Native NTFP trees (kafal, wild pear, oak, Rhododendron) can go into the planting mix as trees. The produce is not the project's to sell, and first fruit takes years.
• Orchard-style tiers (walnut, apple, cardamom understorey), exotic cover crops, resin tapping, fodder lopping and grazing do not belong on the parcel.
• Expect modest, seasonal, unequal NTFP income. In a 24-country household survey, environmental income was 28% of household income, higher in absolute terms for richer households, with the poor relying more on subsistence products (Angelsen et al., 2014). Donor NTFP projects have often failed to deliver, and elite capture is common.
• Cover crops and ground cover: native contour grass and vetiver on gullies and slopes, and mulch rings, are compatible. Their effect on five-year survival in Kumaon is unmeasured. Keep a 0.5–1 m weeded ring around each seedling through April–June.

• Diagnose before planting. Assess every degraded site for ecological memory first. Where residual regenerative capacity exists, release and protect it before replacing it with planted stock.
• Protection and site fit before amendments. Survival in large programmes is explained far more by grazing and fire protection, weeding, tenure and stock size than by what is put in the pit. Match species to aspect, moisture regime and frost risk, and test any amendment before prescribing it.
• Structure serves the credit test and the ecology. Set tree density and species for the 40% canopy test and the surviving-tree count, add native layers for habitat, and report every survival, canopy and biodiversity figure as an estimate with an interval.