DAGRA: Distributed Air-Ground Resilience Architecture

DAGRA structures defence investment into three functionally distinct layers, each with different fragility profiles, adversary targeting logic, and diminishing-returns curves:

The deterrence function is non-additive:

D = (A · G) + (R · (A + G)) − failure_thresholds

This multiplicative structure means A without R is brittle, R without A lacks strategic reach, and G without A cannot prioritise engagements. The three layers are complements, not substitutes — except in catastrophic scenarios where A and G are eliminated and only R determines post-strike survivability.

The diagram below synthesises DRD-06 (bifurcation and hysteresis), DRD-08 (E×D, Black Start, M-component) and DRD-09 (DRI composite) into a single visualisation. Three simulated trajectories show the difference between the 2026 baseline, hardware-only hardening, and the DAGRA 2030 target architecture.

Käyrä A (punainen): Suomi 2026 baseline — romahtaa Faasiin IV T+24h mennessä ilman palautumista. Käyrä B (oranssi katkoviiva): Rauta-optimoitu — hidastaa putoamista mutta jää Faasin III alarajalle. Käyrä C (vihreä): DAGRA 2030 — joustaa Faasiin II, Black Start + NH90 palauttavat koherenssin viikoissa.

The central claim of DRD-08 is not that F-35 is a poor investment in isolation, but that under the 2026 threat environment it represents concentrated vulnerability — eggs in few baskets — while energy and data infrastructure represent a larger structural risk that receives systematically less investment.

The concentration problem

Finland's F-35 fleet of 64 aircraft, even when dispersed under ACE doctrine across highway strips and dispersal bases, remains a finite and enumerable set of high-value targets. Each aircraft represents approximately 150–200 million euros in acquisition cost plus lifecycle overhead. A coordinated first strike with 6–8 Kinzhal missiles against three primary airbases — each capable of Mach 10 penetration at ranges exceeding 2 000 km — would destroy or disable a significant fraction of the fleet before dispersal is complete.

Energy and data infrastructure presents the inverse problem: it is geographically distributed but functionally centralised. Finland's grid has fewer than 20 critical transformer nodes whose simultaneous destruction would black out the country. Fibre backbone routes follow a small number of physical corridors. GPS dependency permeates military and civil systems alike.

The E×D decomposition

R-layer effectiveness is not a single scalar but the product of two independent dimensions:

• E (Energy Resilience Index): probability that critical nodes retain power post-strike. λ_E = 3.0. A small number of HGV strikes against transformer stations can collapse grid over large areas.
• D (Data Resilience Index): probability that C2 communications survive EW and physical attack. λ_D = 4.0. EW at Kinzhal/Tsirkon ranges degrades GPS, radio and fibre simultaneously.

R-layer effective contribution: R_eff = √(E · D)

If either E or D reaches zero, R-layer contribution collapses regardless of investment in the other. A defence posture with excellent energy islanding but no data redundancy — or vice versa — achieves near-zero R-layer effectiveness under combined EW + kinetic attack.

The last row is critical: high energy resilience with low data resilience produces Phase IV collapse. The weakest of E or D governs the system outcome. This has a direct investment implication — the marginal euro should go to whichever of E or D is currently lower, not to the one already higher.

Black Start: the dynamic E-component

Energy resilience (E) has two sub-dimensions. Static E measures whether critical nodes have power during an attack. Dynamic E measures how quickly power is restored after a pulse — the recovery rate γ in DRD-06 Extension C. Black Start capability — automatic inverter-based grid restoration in minutes rather than hours — governs γ directly.

Under pulsed AT(t) with 6-hour inter-pulse intervals, a γ implying 14-hour recovery time means the system cannot return to Phase II coherence before the next pulse. Black Start is therefore not supplementary hardening but a prerequisite for pulsed resilience. Without it, microgrid investment raises static E while leaving dynamic E near zero. Estimated additional cost: ~150–200M EUR above the core energy islanding programme.

Data centre concentration: a new K-term

Rapid data centre growth in southern Finland introduces a new node concentration risk (the K-term in DRD-06 Extension B, C_crit = f(D,R_d,A_q,E_p) − g(K,N_d)). Data centres concentrate grid load and fibre traffic in the area with Finland's highest existing transformer density, creating three compounding effects:

• Shared physical infrastructure: Military C2 nodes and commercial data centres draw from the same transformer stations and fibre corridors. A strike targeting data centre infrastructure simultaneously degrades military C2.
• Increased N_d (network dependency): Smart-grid integration of data centre power management creates correlated failure modes absent in simpler architectures.
• Huoltovarmuus degradation: Supply security experts have warned the data centre boom draws grid capacity toward commercial optimisation at the expense of resilience-critical uses.

DRD-08's R-layer is extended by a third component identified in CN-014 (Distributed Helicopter Resilience, May 2026). The E×D multiplicative structure captures static infrastructure continuity. The M-component captures dynamic, mobile capability that maintains C2, logistics and human rescue even when fixed infrastructure is damaged or destroyed.

The revised R-layer effective contribution:

R_eff = √(E · D) · f(M)

where f(M) is a mobility resilience factor scaling between 0 and 1. At M = 0 (no mobile capability), R_eff collapses regardless of E and D values — static infrastructure without evacuation and logistics continuity cannot sustain C2 under sustained operational stress. At M = 1 (fully integrated dual-use rotary wing assets), R_eff reaches the full √(E·D) value.

Three M-layer functions

• Mobile C2 node: NH90 as a flying command-and-control relay, maintaining situational picture and communications when ground nodes are destroyed or isolated. This directly sustains the D-component under EW and kinetic attack.
• Logistics continuity: Forward fuel, ammunition, spare parts and medical supply delivery when road and rail infrastructure is degraded. This extends the operational timeline before static infrastructure failure forces Phase IV transition.
• Human rescue and evacuation: Civilian evacuation, casualty extraction, HEMS delivery to trauma facilities. Without this, societal endurance (the R-layer floor identified in TN-010) degrades rapidly under sustained pressure.

Current Finnish state: M ≈ 0.2

Finland operates three separate helicopter organisations with no common operational framework: Defence Forces (20 NH90 + 7 MD500, ~30 executive assistance tasks per year), Border Guard (Super Puma, AB206), and FinnHEMS (EC135, H145, 8 bases, civilian HEMS). The ~30 annual executive assistance tasks represent a small fraction of the NH90 fleet's theoretical dual-use capacity. Eastern and northern Finland are underserved by the civilian HEMS network — Rovaniemi is the northernmost base.

Norway's NAWSARH/SAR Queen model (16–18 AW101, six bases, Ministry of Justice ownership, Air Force operation, Joint Rescue Coordination Centre command) achieves M ≈ 0.8 through institutional integration rather than additional procurement. The same aircraft serves both peacetime SAR and wartime military roles. ~1,000 SAR tasks per year versus Finland's ~30 executive assistance tasks — same country size, same population.

Three investment steps to raise M

1. NH90 dual-role systematisation (10–20M EUR one-time): define peacetime SAR/HEMS role and wartime military role; establish cross-ministry steering group (Justice, Defence, Health); equip for HEMS/SAR configuration.
2. HEMS coverage expansion (5–10M EUR/year): position 2–3 NH90 on standby in eastern Finland and Lapland; train crews for paramedic tasks with FinnHEMS; fund from central government, not welfare districts.
3. Reserve specialist model (0–5M EUR/year): target reservist age-limit extension at helicopter operation specialists (aviation maintenance, paramedics, logistics); corporate tax incentive for reserve training funding.

Total budget impact: 10–20M EUR one-time + 5–15M EUR/year. This is <2% of the energy and data infrastructure investment (§03) and <0.1% of F-35 lifecycle cost. The cost-to-strategic-benefit ratio is the highest in the R-layer investment portfolio.

Full analysis: CN-014 — Distributed Helicopter Resilience

Marginal return on R-layer investment

Using the loss function L = Σ w_k · (1 − exp(−λ_k · α_k)) with weights w_A=0.5, w_R=0.3, w_G=0.2, sensitivity analysis compares the effect of a marginal 100M EUR added to A-layer versus R-layer:

R-layer marginal returns exceed A-layer returns by 5–15× when R < 25% of budget. Finland's current R allocation (~10%) places it in the 15× marginal return zone — the highest possible infrastructure investment leverage point.

Critical operational thresholds

Phase II C2 coherence (CI ≥ 0.6) requires E ≥ 0.5 and D ≥ 0.5 simultaneously. Estimated current Finnish state: E ≈ 0.3–0.4, D ≈ 0.4–0.5. This places the system in Phase III under moderate operational stress — without any kinetic engagement.

~15–20% of annual defence budget as one-time investment. Transition from estimated current state (E≈0.35, D≈0.40) to Phase II threshold (E≥0.50, D≥0.50). All three categories are required — the E×D multiplicative structure means partial investment in only one dimension produces near-zero R-layer gain.

Against Avangard, A-layer and G-layer are operationally irrelevant as defensive instruments. The only structurally coherent responses are NATO collective deterrence (Article 5) and R-layer post-strike survivability. This is the strongest possible argument for R-layer investment: it is the sole defensive layer with relevance across the full threat spectrum from drone swarms through conventional hypersonics to nuclear HGV.

Note on production constraints: Russia has reported Avangard production limitations. Small inventory does not reduce strategic threat — a handful of strikes against Finnish C2 and energy nodes would achieve rapid systemic collapse. Limited inventory reinforces rather than diminishes the R-hardening case.

Path dependency note: the F-35 lifecycle commitment (~25–30B EUR over 30 years, with ~730M EUR currency losses already accrued) constrains rapid reallocation. The realistic transition path is 60/20/10 → 45/35/20 (near-term) → 40/35/25 (medium-term as F-35 early costs are absorbed). The critical constraint is that R must reach 20% before G reaches 35% — the multiplicative E×D structure means an under-resourced R negates G investment effectiveness in extended conflict.

1. Energy islanding of C2 critical nodes — highest marginal return (15× vs A-layer), achieves E ≥ 0.5 threshold. ~500M EUR. Timeline: 3–5 years.
2. Data redundancy infrastructure — satellite terminals, HF radio, fibre rerouting. Achieves D ≥ 0.5. ~300M EUR. Timeline: 2–4 years.
3. GBAD expansion — Kinzhal is the primary design case. NASAMS/Patriot mobile units, C-UAS. ~800M EUR/yr additional. Ongoing.
4. F-35 dispersal infrastructure — ACE highway strips, hardened dispersal points. Reduces A-layer concentration without reducing airframe count. ~300M EUR.
5. Further F-35 acquisition — lowest marginal return once R ≥ 25% and G ≥ 35%. Lifecycle commitments already made; future acquisition decisions should apply diminishing-returns test explicitly.

The R-layer investment priorities in §06 have concrete implementation pathways in the ACI research programme — operationalised frameworks addressing the same structural gap from the civilian energy system side.

SGFA and MESA were developed as civilian energy resilience instruments. DRD-08's R-layer requirements and the SGFA/MESA implementation pathways are converging solutions to the same structural problem: Finland's critical infrastructure is optimised for efficiency under stable conditions and fragile under simultaneous grid stress and adversarial pressure. The investment case is identical from either direction.