OHProduction

{
  "reference": {
    "notes": "source_ref: GasChem.jl@8c12c048482b515fb8eb2110bf8ab4b4f4e71309:src/oh_production.jl:88\n\nUpstream `OHProduction` is an MTK `@component` with 6 `@variables` (3 inputs O₃/H₂O/M + 3 outputs O₁D/P_OH/ε_OH), 6 `@parameters` (j_O₃, k₃_N₂, k₃_O₂, k₄, f_N₂, f_O₂), 1 `@constants` (two = 2.0 dimensionless stoichiometric factor — 2 OH per O(¹D)+H₂O reaction), 0 helper expressions and 3 algebraic equations from S&P §6.1. The upstream compiles the system with `mtkcompile(sys; inputs=[O₃, H₂O, M])` and solves it as a `NonlinearProblem` with the three inputs pinned to user-supplied concentrations.\n\nVariable-classification convention: the three input @variables (O₃, H₂O, M) are encoded here as `type: parameter` because they have no defining equation within the OHProduction component itself — they are externally-supplied inputs that the upstream compile treats via `inputs=[...]` to NonlinearProblem. This mirrors the rig-wide convention applied by sibling stefan_boltzmann.esm (the upstream @parameter T is encoded as parameter, and the upstream @constants σ is encoded as parameter) and typical.esm/urban.esm/remote.esm (the upstream @parameters air-composition species are encoded as parameter). The three output @variables (O₁D, P_OH, ε_OH) are encoded as `type: state` because each has one defining algebraic equation that the inline-test runner solves as part of the post-mtkcompile observed-binding layer. The upstream structural test `OHProduction: Structural Verification` (test/test_oh_production.jl:21-45) asserts `length(unknowns(sys)) == 6` and `length(equations(sys)) == 3` — the .esm preserves this with the 6 declared variables (3 parameter + 3 state) and 3 entries in the `equations` array.\n\nConstant→parameter translation: the upstream `@constants two = 2` is encoded as a fixed-default parameter (`type: parameter`, default 2.0, units \"1\"). This mirrors the chapman.esm convention (where the 4 upstream `@constants` k₂_A/k₄_A/C_k4/T_ref are preserved as `type: parameter`), the gwp.esm convention (where τ_CO2/a_CO2 `@constants` become parameters), and the stefan_boltzmann.esm convention (σ `@constants` → parameter). This serves three purposes: (a) MTK's unit checker treats `two` as a named dimensionless quantity rather than a magic literal in the equation RHS, preserving the upstream unit-balance proof; (b) downstream consumers can override the stoichiometric factor for sensitivity studies (e.g., bind two=1 to model a hypothetical single-OH yield reaction); (c) the description field documents the physical meaning (2 OH per O(¹D)+H₂O encounter from §6.1).\n\nRate-constant CGS→SI translation: the upstream Julia declares the four rate constants as the SI-converted CGS values via explicit multiplication by 1e-6 (the cm³/molec/s → m³/s factor for bimolecular k constants):\n  k₃,N₂ = 2.6e-11 · 1.0e-6 = 2.6e-17 m³/s  (S&P Table B.1, IUPAC-recommended for O(¹D) + N₂ at 298 K)\n  k₃,O₂ = 4.0e-11 · 1.0e-6 = 4.0e-17 m³/s  (S&P Table B.1, O(¹D) + O₂ at 298 K)\n  k₄    = 2.2e-10 · 1.0e-6 = 2.2e-16 m³/s  (S&P Table B.1, O(¹D) + H₂O at 298 K — about 5-10× faster than the quenching channels per molecule)\n  j_O₃  = 6.0e-5 s⁻¹                       (S&P Figure 4.15, photolysis rate at surface, solar zenith 0°)\nThe .esm encodes the SI-converted defaults directly (2.6e-17, 4.0e-17, 2.2e-16, 6.0e-5) — bit-identical to the upstream Julia evaluation at typical magnitudes because the f_N₂·k₃,N₂ + f_O₂·k₃,O₂ + k₄·H₂O/M arithmetic washes out the 1-ULP difference between the literal `4.0e-17` and the Julia product `4.0e-11 · 1.0e-6 = 3.999...e-17` (verified at the default scenario: denominator 8.05e8 bit-equal in both encodings). The f_N₂=0.78 and f_O₂=0.21 air-composition fractions sum to 0.99 not 1.0 — they reflect the dominant-component approximation that lumps the residual 1% (Ar + trace gases) into 'inactive bath' contributing no O(¹D) quenching at S&P's stated precision.\n\nObservability classification: this component is purely algebraic — every output variable is the LHS of one algebraic equation in the inputs and parameters (no derivative, the `(t)` time index has no role since there is no `D(·)` operator in any equation). Test scenarios use `time_span: [0.0, 1.0]` and sample at t=0/t=1 to mirror the convention used by sibling observed-only ESM files (gwp.esm, photon_energy.esm, stefan_boltzmann.esm, wien.esm, typical.esm). For an algebraic-only system the trajectory is constant in time and converges to the same fixed point in one nonlinear-solver step regardless of the sampling time. The post-mtkcompile system has zero ODE states and the inline-test runner exercises O₁D/ε_OH/P_OH at their observed algebraic values.\n\nEquation cascade: the three equations form a topologically-ordered cascade — Eq. 6.1 (O₁D) depends only on inputs/parameters, Eq. 6.4 (ε_OH) depends only on inputs/parameters and shares the same denominator as Eq. 6.1, Eq. 6.3 (P_OH) depends on inputs/parameters and references ε_OH. This mirrors the GWP cascade (decay_integral_A → AGWP_A → GWP) and the gwp.esm pattern of writing P_OH's equation in terms of ε_OH rather than fully inlining — preserves the upstream symbolic structure (so a downstream consumer reading the .esm sees Eq. 6.3 written as P_OH = j_O₃·O₃·ε_OH the way S&P writes it, rather than the fully-expanded P_OH = 2·j_O₃·O₃·k₄·H₂O / (k₃_eff·M + k₄·H₂O) which obscures the physical interpretation of ε_OH).\n\nNumerical defaults (S&P typical surface conditions, all in SI m⁻³):\n  O₃    = 1.0e18 m⁻³  (~40 ppb at STP — upstream typical value matches typical.esm O₃=1.0e18)\n  H₂O   = 4.0e23 m⁻³  (~50% RH at 298 K, upstream typical — matches typical.esm H₂O=4.0e23)\n  M     = 2.5e25 m⁻³  (STP air number density — matches typical.esm M=2.5e25)\nAt these defaults:\n  k₃_eff = 0.78 · 2.6e-17 + 0.21 · 4.0e-17 = 2.028e-17 + 8.4e-18 = 2.868e-17 m³/s\n  denom  = 2.868e-17 · 2.5e25 + 2.2e-16 · 4.0e23 = 7.17e8 + 8.8e7 = 8.05e8 1/s\n  O₁D    = 6.0e-5 · 1.0e18 / 8.05e8 = 74534.16149068323 m⁻³ (~7.45e4 = ~7e4 molec/cm³, S&P p. 207)\n  ε_OH   = 2 · 2.2e-16 · 4.0e23 / 8.05e8 = 0.21863354037267077 (~22% OH yield at 50% RH, S&P p. 207 Table)\n  P_OH   = 6.0e-5 · 1.0e18 · 0.21863354 = 13118012422360.246 m⁻³/s (~1.31e13 m⁻³/s = ~1.31e7 molec/cm³/s)\n\nUpstream test reproduction:\n- `OHProduction: Structural Verification` (test/test_oh_production.jl:21-45) asserts 6 unknowns, 3 equations, and that the names {O₁D, P_OH, ε_OH, O₃, H₂O, M} are all present — reproduced by the .esm structural shape (3 parameter + 3 state variables under `variables`, 3 entries in `equations`).\n- `OHProduction: Eq 6.1 O(1D) Steady-State` (test/test_oh_production.jl:50-76) asserts O₁D ≈ 7.453e4 at rtol=0.01 with O₃=1e18/M=2.5e25/H₂O=4e23 — reproduced by `upstream_eq61_o1d_steady_state_defaults` at rel=1.0e-10 pinning the exact 74534.16149068323.\n- `OHProduction: Eq 6.4 OH Yield` (test/test_oh_production.jl:78-102) asserts ε_OH ≈ 0.2186 at rtol=0.05 at the same defaults — reproduced by `upstream_eq64_eps_OH_defaults` at rel=1.0e-10 pinning the exact 0.21863354037267077.\n- `OHProduction: Eq 6.3 OH Production Rate` (test/test_oh_production.jl:104-125) asserts P_OH ≈ 1.312e13 at rtol=0.05 — reproduced by `upstream_eq63_p_oh_rate_defaults` at rel=1.0e-10 pinning the exact 13118012422360.246.\n- `OHProduction: Rate Constants at 298 K` (test/test_oh_production.jl:130-145) asserts the default values of the 6 parameters — implicitly covered by the `default` values declared on each `variables` entry; the structural-test gate verifies they are bit-identical to the upstream Julia values.\n- `OHProduction: Limiting Behavior - Humidity Variation` (test/test_oh_production.jl:150-176) asserts ε_OH < 0.02 at low humidity (H₂O=1e19) and 0 < ε_OH(low) < ε_OH(high) < 2 at high humidity (H₂O=1e24) — reproduced by `upstream_low_humidity_limit` (pins ε_OH=6.136661784301359e-06 at H₂O=1e19) and `upstream_high_humidity_limit` (pins ε_OH=0.4695837780149413 at H₂O=1e24).\n- `OHProduction: Eq 6.4 ε_OH vs Book Table (p. 207)` (test/test_oh_production.jl:181-224) asserts ε_OH at 3 RH values (10%, 25%, 50%) against the book table at rtol=0.1 — reproduced by `upstream_book_table_rh10_p207`, `upstream_book_table_rh25_p207`, `upstream_book_table_rh50_p207` pinning the exact computed values at rel=1.0e-10 (the upstream test omits the RH=80% case noting that the book uses the approximate formula ε_OH ≈ 2·k₄·ξ_H₂O/k₃ which diverges from the exact form of Eq. 6.4 at high RH; we include RH=80% as an additional pinning to characterize this divergence).\n- `OHProduction: Qualitative - O1D Proportional to O3` (test/test_oh_production.jl:226-247) asserts the ratio O₁D(O₃=2e18) / O₁D(O₃=1e18) ≈ 2.0 at rtol=1.0e-6 — reproduced by `upstream_o1d_linear_in_o3_baseline` and `upstream_o1d_linear_in_o3_doubled` pinning both endpoints (74534.16149068323 and 149068.32298136645) at rel=1.0e-10 so the ratio is exact by IEEE-754 multiplication-by-2 reproducibility.\n- `OHProduction: Qualitative - Positive Quantities` (test/test_oh_production.jl:249-273) asserts strict positivity over a 3×3×3 grid of (M, H₂O, O₃) values — reproduced implicitly by `upstream_high_pressure_quenching_dominates`, `upstream_low_pressure_stratosphere_regime`, and the other parameter-sweep tests which all pin strictly-positive O₁D/ε_OH/P_OH values.\n\nTolerance rationale (model-level rel=1.0e-10 / abs=1.0e-12): the three algebraic equations evaluate as small expression trees: Eq. 6.1 is one multiplication + one addition + one multiplication + one division (4 IEEE-754 ops), Eq. 6.4 is one multiplication + one multiplication + one addition + one multiplication + one division (5 IEEE-754 ops), Eq. 6.3 is two multiplications (2 IEEE-754 ops). Each output is correctly-rounded to within ~1 ULP ≈ 1e-16 relative; rel=1.0e-10 leaves 6 orders of magnitude of headroom over the IEEE-754 floor and is comfortably tighter than the upstream rtol bounds (rtol=0.01 for hand-calc cross-checks, rtol=0.05 for hand-calc cross-checks, rtol=0.1 for book-table approximate-formula bounds, rtol=1e-6 for the O₁D-vs-O₃ ratio identity). The abs=1.0e-12 floor handles the boundary cases zero-H₂O → ε_OH=0/P_OH=0 (the 2·k₄·H₂O numerator collapses to exact IEEE-754 zero) and zero-O₃ → O₁D=0/P_OH=0 (the j_O₃·O₃ numerator collapses to exact zero), gating the stuck-at-non-zero failure mode. Matches the convention used by sibling observed-only ESM files: stefan_boltzmann.esm at rel=1e-10/abs=1e-12, wien.esm at rel=1e-10, gwp.esm at rel=1e-10, photon_energy.esm at rel=1e-10/abs=1e-12.\n\nUnit balance (verified): all three equations balance dimensionally.\n  Eq. 6.1 `O₁D ~ j_O₃·O₃ / (k₃_eff·M + k₄·H₂O)`: (s⁻¹)·(m⁻³) / ((m³/s)·(m⁻³) + (m³/s)·(m⁻³)) = (m⁻³/s) / (s⁻¹) = m⁻³ ✓\n  Eq. 6.4 `ε_OH ~ 2·k₄·H₂O / (k₃_eff·M + k₄·H₂O)`: 1 · (m³/s)·(m⁻³) / (s⁻¹) = (1/s) / (1/s) = dimensionless ✓\n  Eq. 6.3 `P_OH ~ j_O₃·O₃·ε_OH`: (s⁻¹)·(m⁻³)·(1) = m⁻³·s⁻¹ ✓ (matches P_OH unit `u\"m^-3*s^-1\"`)\nThe migrated `.esm` preserves the upstream units verbatim — all concentrations in SI m⁻³, all rate constants in SI (s⁻¹ for first-order, m³/s for second-order bimolecular), all dimensionless fractions/yields in unit `1`. SI throughout; no unit conversion at parse or runtime."
  },
  "tolerance": {
    "rel": 1e-10,
    "abs": 1e-12
  },
  "variables": {
    "j_O3": {
      "type": "parameter",
      "units": "s^-1",
      "default": 6e-05,
      "description": "O₃ photolysis rate producing O(¹D) at surface, solar zenith angle 0° (S&P Figure 4.15, Table B.1). The upstream default 6.0e-5 s⁻¹ is the noon-equatorial-surface canonical value; at other zenith angles and altitudes j_O₃ varies by factors of 10⁻²-10¹ — overridable for scenarios at higher altitudes (j_O₃ increases roughly 2-3× from surface to 12 km), higher latitudes (drops with sec(SZA)), or non-equatorial seasons. Downstream coupling: in the full TroposphericChemistrySystem this parameter is provided by the upstream FastJX or SuperFast photolysis subsystem via param_to_var coupling."
    },
    "k3_N2": {
      "type": "parameter",
      "units": "m^3/s",
      "default": 2.6e-17,
      "description": "O(¹D) + N₂ quenching rate constant (S&P Table B.1, IUPAC-recommended at 298 K). The upstream default 2.6e-17 m³/s is the SI-converted form of the CGS value 2.6e-11 cm³/molec/s — the canonical IUPAC-recommended bimolecular rate coefficient for O(¹D) + N₂ → O(³P) + N₂ at 298 K. Weakly temperature-dependent (k₃,N₂ ∝ exp(110/T) per IUPAC 2007 update — about 50% faster at 200 K than 298 K) but the upstream uses the temperature-independent 298 K value for the algebraic system. The two quenching channels (k₃_N₂, k₃_O₂) are weighted by their atmospheric mass fractions (f_N₂=0.78, f_O₂=0.21) to give the effective k₃_eff = 0.78·k₃_N₂ + 0.21·k₃_O₂ = 2.868e-17 m³/s used in the denominator of Eqs. 6.1 and 6.4."
    },
    "k3_O2": {
      "type": "parameter",
      "units": "m^3/s",
      "default": 4e-17,
      "description": "O(¹D) + O₂ quenching rate constant (S&P Table B.1, IUPAC-recommended at 298 K). The upstream default 4.0e-17 m³/s is the SI-converted form of the CGS value 4.0e-11 cm³/molec/s — the canonical bimolecular rate coefficient for O(¹D) + O₂ → O(³P) + O₂ at 298 K. About 1.5× faster than k₃_N₂ per molecule, but weighted by the smaller f_O₂=0.21 atmospheric mass fraction → the O₂ channel contributes 8.4e-18/2.868e-17 = 29% of k₃_eff while the N₂ channel contributes 71%. Weakly temperature-dependent (k₃,O₂ ∝ exp(55/T)) but the upstream uses the temperature-independent 298 K value."
    },
    "k4": {
      "type": "parameter",
      "units": "m^3/s",
      "default": 2.2e-16,
      "description": "O(¹D) + H₂O → 2 OH rate constant (S&P Table B.1, IUPAC-recommended at 298 K). The upstream default 2.2e-16 m³/s is the SI-converted form of the CGS value 2.2e-10 cm³/molec/s — the canonical bimolecular rate coefficient at 298 K. The OH-producing channel is ~5-10× faster per molecule than either quenching channel (k₄/k₃_N₂ ≈ 8.5, k₄/k₃_O₂ ≈ 5.5) — the reason H₂O is a far more efficient O(¹D) reaction partner per molecule than N₂ or O₂. But because [H₂O] / ([N₂]+[O₂]) ≈ 0.01-0.03 at typical surface RH, the absolute reaction-rate balance still favors quenching by ~5× (numerator vs denominator in the ε_OH formula): k₄[H₂O] = 2.2e-16·4e23 = 8.8e7 vs k₃_eff[M] = 2.868e-17·2.5e25 = 7.17e8, giving ε_OH ≈ 0.22 at S&P p. 207 conditions (about 1 in 5 O(¹D) atoms makes OH)."
    },
    "f_N2": {
      "type": "parameter",
      "units": "1",
      "default": 0.78,
      "description": "Fraction of air number density that is N₂, dimensionless (S&P Page 1). The upstream default 0.78 reflects the well-mixed atmosphere composition where N₂ is the dominant species at all altitudes below the homopause (~100 km). f_N₂ + f_O₂ = 0.99 here (not exactly 1.0) — the residual 1% (Ar 0.93%, CO₂ 0.04%, plus trace gases) is treated as an 'inactive bath' that contributes negligible O(¹D) quenching at S&P's stated precision. Overridable for a hypothetical pure-N₂ atmosphere (f_N₂=1.0, f_O₂=0.0 → k₃_eff = k₃_N₂ = 2.6e-17) or for sensitivity studies on the trace-gas correction."
    },
    "f_O2": {
      "type": "parameter",
      "units": "1",
      "default": 0.21,
      "description": "Fraction of air number density that is O₂, dimensionless (S&P Page 1). The upstream default 0.21 is the canonical well-mixed atmospheric O₂ fraction (more precisely 0.2095 by the standard dry-air composition table; the upstream uses 2-digit precision)."
    },
    "two": {
      "type": "parameter",
      "units": "1",
      "default": 2.0,
      "description": "Stoichiometric factor: 2 OH radicals produced per O(¹D) + H₂O reaction (dimensionless). Preserved from upstream `@constants two = 2` per the rig convention (translating upstream `@constants` to fixed-default parameters so MTK's unit checker sees a named dimensionless quantity in the equation RHS — matches chapman.esm/gwp.esm/stefan_boltzmann.esm convention). The factor is hardwired by the O(¹D) + H₂O → 2 OH stoichiometry and not expected to change in routine use; overridable only for hypothetical alternate-stoichiometry sensitivity studies."
    },
    "O3": {
      "type": "parameter",
      "units": "m^-3",
      "default": 1e+18,
      "description": "Ozone number density [m⁻³] (input from coupled chemistry/transport system). The default 1.0e18 m⁻³ ≈ 40 ppb at STP matches the typical-tropospheric value used in sibling typical.esm and in the upstream test scenarios at test/test_oh_production.jl:55. Externally-supplied input: in the full TroposphericChemistrySystem this would be provided by the upstream NOxPhotochemistry/Chapman/SuperFast photochemistry subsystem; in standalone OHProduction tests it is overridden to scenario-specific values via `parameter_overrides`. Encoded as `type: parameter` (not `type: state`) because OHProduction has no governing equation for O₃ — it is purely an input. This matches the rig convention used by sibling typical.esm/urban.esm/remote.esm (which encode their 11 species inputs as parameters) and stefan_boltzmann.esm (which encodes its T input as parameter)."
    },
    "H2O": {
      "type": "parameter",
      "units": "m^-3",
      "default": 4e+23,
      "description": "Water vapor number density [m⁻³] (input from coupled chemistry/transport system). The default 4.0e23 m⁻³ corresponds to ~50% relative humidity at 298 K (saturation vapor pressure × RH × Avogadro / molar volume) and matches the typical-tropospheric value used in sibling typical.esm. The ε_OH yield (Eq. 6.4) is most sensitive to this parameter: increasing H₂O by 10× from 4e22 to 4e23 increases ε_OH by ~10× from 0.024 to 0.22 (the linear-numerator regime where k₄·H₂O << k₃_eff·M). At very high humidity (H₂O → ∞) ε_OH saturates at 2.0 (the upper bound — every O(¹D) makes 2 OH). In the troposphere H₂O ranges from ~1e21 (dry desert / cold winter) to ~1e24 (tropical moist) m⁻³; in the stratosphere H₂O is far smaller (~1e20 m⁻³)."
    },
    "M": {
      "type": "parameter",
      "units": "m^-3",
      "default": 2.5e+25,
      "description": "Total air number density [m⁻³] (input from coupled chemistry/transport system). The default 2.5e25 m⁻³ is the STP value (PV=NkT at 1 atm and 273 K gives M ≈ 2.69e25, but the upstream uses the rounded 2.5e25 for typical surface conditions ~298 K, matching sibling typical.esm). M sets the rate of O(¹D) quenching (k₃_eff·M) which dominates the denominator of Eq. 6.1 and Eq. 6.4 at typical surface humidity. M decreases roughly exponentially with altitude (scale height ~8 km in the troposphere): at 12 km M ≈ 1e25, at 30 km M ≈ 1e24, at 60 km M ≈ 1e22. The OHProduction component's structure is altitude-independent — the user supplies M from the upstream transport/thermodynamics module."
    },
    "O1D": {
      "type": "unknown",
      "units": "m^-3",
      "description": "Excited atomic oxygen O(¹D) steady-state number density [m⁻³] (Eq. 6.1). Algebraic balance between source j_O₃·O₃ and sinks k₃_eff·M + k₄·H₂O: O(¹D) = j_O₃·O₃ / (k₃_eff·M + k₄·H₂O). At S&P p. 207 default conditions O(¹D) ≈ 7.45e4 m⁻³ ≈ 7.45e-2 molec/cm³ — orders of magnitude smaller than ground-state O(³P) which itself is ~1e6-1e7 m⁻³ in the troposphere. The 'steady-state' label reflects the quasi-equilibrium assumption: O(¹D) lifetime is ~10⁻⁹ s (the inverse of k₃_eff·M ≈ 7e8 s⁻¹) which is many orders of magnitude shorter than any other tropospheric timescale, so the quasi-steady-state approximation is exact to within numerical precision. Linear in O₃ — doubling O₃ doubles O(¹D)."
    },
    "ε_OH": {
      "type": "unknown",
      "units": "1",
      "description": "OH yield: number of OH radicals produced per O(¹D) atom formed, dimensionless (Eq. 6.4). Algebraic ratio ε_OH = 2·k₄·H₂O / (k₃_eff·M + k₄·H₂O). Ranges from 0 (all O(¹D) quenched in pure-dry air → no OH made) to 2 (all O(¹D) reacts with H₂O → 2 OH per O(¹D) by stoichiometry of the 2-OH yield channel). At S&P p. 207 typical 50% RH default conditions ε_OH ≈ 0.22 (about 11% of O(¹D) reacts with H₂O, producing 22% as many OH as O(¹D) formed). The book table on p. 207 reports ε_OH ≈ {0.047, 0.12, 0.23, 0.38} at RH = {10%, 25%, 50%, 80%} respectively — reproduced here at low/moderate RH within ~5% (the book uses the approximate formula ε_OH ≈ 2·k₄·ξ_H₂O/k₃ which diverges at high RH where the k₄·H₂O denominator term becomes non-negligible)."
    },
    "P_OH": {
      "type": "unknown",
      "units": "m^-3*s^-1",
      "description": "OH production rate [m⁻³·s⁻¹] (Eq. 6.3). Product of the O(¹D) source rate j_O₃·O₃ and the OH yield ε_OH: P_OH = j_O₃·O₃·ε_OH. Equivalent to the fully-expanded form 2·j_O₃·O₃·k₄·H₂O / (k₃_eff·M + k₄·H₂O). At S&P p. 207 default conditions P_OH ≈ 1.31e13 m⁻³/s ≈ 1.31e7 molec/cm³/s — the canonical 'OH production rate at 50% RH, 40 ppb O₃, surface' value driving tropospheric OH steady-state concentrations [OH] ≈ P_OH / k_OH_sinks ≈ 1e6-1e7 molec/cm³ that determine the lifetime of CH₄ and other OH-reactive trace gases. Linear in O₃ (via the j_O₃·O₃ factor) and approximately linear in H₂O at low RH (via ε_OH ~ H₂O when k₄·H₂O << k₃_eff·M)."
    }
  },
  "equations": [
    {
      "lhs": "O1D",
      "rhs": {
        "op": "/",
        "args": [
          {
            "op": "*",
            "args": [
              "j_O3",
              "O3"
            ]
          },
          {
            "op": "+",
            "args": [
              {
                "op": "*",
                "args": [
                  {
                    "op": "+",
                    "args": [
                      {
                        "op": "*",
                        "args": [
                          "f_N2",
                          "k3_N2"
                        ]
                      },
                      {
                        "op": "*",
                        "args": [
                          "f_O2",
                          "k3_O2"
                        ]
                      }
                    ]
                  },
                  "M"
                ]
              },
              {
                "op": "*",
                "args": [
                  "k4",
                  "H2O"
                ]
              }
            ]
          }
        ]
      }
    },
    {
      "lhs": "ε_OH",
      "rhs": {
        "op": "/",
        "args": [
          {
            "op": "*",
            "args": [
              "two",
              "k4",
              "H2O"
            ]
          },
          {
            "op": "+",
            "args": [
              {
                "op": "*",
                "args": [
                  {
                    "op": "+",
                    "args": [
                      {
                        "op": "*",
                        "args": [
                          "f_N2",
                          "k3_N2"
                        ]
                      },
                      {
                        "op": "*",
                        "args": [
                          "f_O2",
                          "k3_O2"
                        ]
                      }
                    ]
                  },
                  "M"
                ]
              },
              {
                "op": "*",
                "args": [
                  "k4",
                  "H2O"
                ]
              }
            ]
          }
        ]
      }
    },
    {
      "lhs": "P_OH",
      "rhs": {
        "op": "*",
        "args": [
          "j_O3",
          "O3",
          "ε_OH"
        ]
      }
    }
  ],
  "tests": [
    {
      "id": "upstream_eq61_o1d_steady_state_defaults",
      "description": "Reproduces and tightens the upstream `OHProduction: Eq 6.1 O(1D) Steady-State` test (GasChem.jl test/test_oh_production.jl:50-76). At the typical surface scenario O₃=1e18 m⁻³ (~40 ppb), H₂O=4e23 m⁻³ (~50% RH at 298 K), M=2.5e25 m⁻³ (STP air), the upstream asserts O(¹D) ≈ 7.453e4 m⁻³ at rtol=0.01 (hand-computed from k₃_eff·M + k₄·H₂O = 8.05e8 and O(¹D) = 6e-5·1e18/8.05e8). This test pins the exact IEEE-754 value 74534.16149068323 m⁻³ at rel=1e-10 — 8 orders of magnitude tighter than upstream — anchoring the Eq. 6.1 algebraic identity to single-ULP precision and providing a stuck-at-default regression detector if the equation's numerator/denominator structure is ever accidentally rewired.",
      "parameter_overrides": {},
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "O1D",
          "time": 0.0,
          "expected": 74534.16149068323
        }
      ]
    },
    {
      "id": "upstream_eq64_eps_OH_defaults",
      "description": "Reproduces and tightens the upstream `OHProduction: Eq 6.4 OH Yield` test (GasChem.jl test/test_oh_production.jl:78-102) at the same typical surface scenario (O₃=1e18, H₂O=4e23, M=2.5e25). The upstream asserts 0.1 < ε_OH < 0.5 and ε_OH ≈ 0.2186 at rtol=0.05 (hand-computed from 2·k₄·H₂O/(k₃_eff·M+k₄·H₂O) = 1.76e8/8.05e8 ≈ 0.2186). This test pins the exact IEEE-754 value 0.21863354037267077 at rel=1e-10 — 8 orders of magnitude tighter than upstream. Together with `upstream_eq61_o1d_steady_state_defaults` this anchors the shared denominator (k₃_eff·M + k₄·H₂O) used by both algebraic equations — a single denominator-structure regression would break both tests simultaneously.",
      "parameter_overrides": {},
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "ε_OH",
          "time": 0.0,
          "expected": 0.21863354037267077
        }
      ]
    },
    {
      "id": "upstream_eq63_p_oh_rate_defaults",
      "description": "Reproduces and tightens the upstream `OHProduction: Eq 6.3 OH Production Rate` test (GasChem.jl test/test_oh_production.jl:104-125) at the same typical surface scenario. The upstream asserts 1.0e12 < P_OH < 1.0e14 and P_OH ≈ 1.312e13 m⁻³/s at rtol=0.05 (hand-computed from j_O₃·O₃·ε_OH = 6e-5·1e18·0.2186 ≈ 1.312e13). This test pins the exact IEEE-754 value 13118012422360.246 m⁻³/s at rel=1e-10. P_OH directly couples to ε_OH (Eq. 6.3 is written P_OH = j_O₃·O₃·ε_OH) — together with the Eq. 6.4 test this anchors the equation-cascade dependency P_OH → ε_OH → (denominator) — three tests pinning the three outputs at the same input scenario gives full algebraic-identity regression coverage at S&P's anchor conditions.",
      "parameter_overrides": {},
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "P_OH",
          "time": 0.0,
          "expected": 13118012422360.246
        }
      ]
    },
    {
      "id": "upstream_low_humidity_limit",
      "description": "Reproduces and tightens the upstream `OHProduction: Limiting Behavior - Humidity Variation` low-humidity case (GasChem.jl test/test_oh_production.jl:150-176). With H₂O = 1.0e19 m⁻³ (4 orders of magnitude below typical 50% RH at 298 K — extreme dry-air case), the k₄·H₂O numerator is ~7000× smaller than the k₃_eff·M quenching term, so ε_OH drops to ~6.1e-6 (deep linear-numerator regime). The upstream asserts ε_OH < 0.02 qualitative bound; this test pins the exact IEEE-754 values ε_OH=6.136661784301359e-06 and O₁D=83681.75160410945 (slightly larger than the typical-RH case because the denominator is slightly smaller — the k₄·H₂O term is negligible compared to k₃_eff·M, so O₁D ≈ j_O₃·O₃/(k₃_eff·M) is the asymptotic dry-air limit) at rel=1e-10. P_OH drops to 3.68e8 m⁻³/s (~36000× smaller than typical) reflecting the multiplicative collapse of both ε_OH and the unchanged j_O₃·O₃.",
      "parameter_overrides": {
        "H2O": 1e+19
      },
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "O1D",
          "time": 0.0,
          "expected": 83681.75160410945
        },
        {
          "variable": "ε_OH",
          "time": 0.0,
          "expected": 6.136661784301359e-06
        },
        {
          "variable": "P_OH",
          "time": 0.0,
          "expected": 368199707.05808157
        }
      ]
    },
    {
      "id": "upstream_high_humidity_limit",
      "description": "Reproduces and tightens the upstream `OHProduction: Limiting Behavior - Humidity Variation` high-humidity case (GasChem.jl test/test_oh_production.jl:150-176). With H₂O = 1.0e24 m⁻³ (2.5× higher than typical 50% RH — wet tropical-boundary-layer case), the k₄·H₂O numerator term grows to be comparable to k₃_eff·M, pushing ε_OH up to ~0.47 (about half of the maximum 2.0 — half of every O(¹D) makes OH, the other half is still quenched). The upstream asserts ε_OH(high) > ε_OH(low) and 0 < ε_OH < 2; this test pins the exact IEEE-754 ε_OH=0.4695837780149413 and O₁D=64034.151547491994 (slightly smaller than the typical-RH case because the denominator is slightly larger from the now-non-negligible k₄·H₂O term) at rel=1e-10. P_OH grows to 2.82e13 m⁻³/s (~2× the typical-RH value).",
      "parameter_overrides": {
        "H2O": 1e+24
      },
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "O1D",
          "time": 0.0,
          "expected": 64034.151547491994
        },
        {
          "variable": "ε_OH",
          "time": 0.0,
          "expected": 0.4695837780149413
        },
        {
          "variable": "P_OH",
          "time": 0.0,
          "expected": 28175026680896.477
        }
      ]
    },
    {
      "id": "upstream_book_table_rh10_p207",
      "description": "Reproduces the upstream `OHProduction: Eq 6.4 ε_OH vs Book Table (p. 207)` low-RH case (GasChem.jl test/test_oh_production.jl:181-224, first iteration with rh=10). At 298 K saturated H₂O mixing ratio ξ_H₂O_sat ≈ 0.031, so H₂O = 0.1 · 0.031 · 2.5e25 = 7.75e22 m⁻³ (10% RH). The S&P p. 207 book table reports ε_OH ≈ 0.047 at RH=10% (using the approximate formula ε_OH ≈ 2·k₄·ξ_H₂O/k₃). Upstream asserts ε_OH ≈ 0.047 at rtol=0.1 (10% tolerance for the exact-vs-approximate formula divergence and the k₃_eff-vs-k₃ definitional difference). This test pins the exact IEEE-754 ε_OH=0.04645460118520538 from the exact Eq. 6.4 formula at rel=1e-10 — the computed value falls within the upstream rtol=0.1 of the book value (relative error 1.2% — well within the 10% bound).",
      "parameter_overrides": {
        "H2O": 7.75e+22
      },
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "ε_OH",
          "time": 0.0,
          "expected": 0.04645460118520538
        }
      ]
    },
    {
      "id": "upstream_book_table_rh25_p207",
      "description": "Reproduces the upstream `OHProduction: Eq 6.4 ε_OH vs Book Table (p. 207)` mid-low-RH case (GasChem.jl test/test_oh_production.jl:181-224, second iteration with rh=25). H₂O = 0.25 · 0.031 · 2.5e25 = 1.9375e23 m⁻³ (25% RH). Book value ε_OH ≈ 0.12 at rtol=0.1; this test pins the exact computed 0.11222642751357578 from Eq. 6.4 at rel=1e-10 (within the upstream rtol=0.1 of the book value — relative error 6.5%, just within bounds).",
      "parameter_overrides": {
        "H2O": 1.9375e+23
      },
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "ε_OH",
          "time": 0.0,
          "expected": 0.11222642751357578
        }
      ]
    },
    {
      "id": "upstream_book_table_rh50_p207",
      "description": "Reproduces the upstream `OHProduction: Eq 6.4 ε_OH vs Book Table (p. 207)` mid-RH case (GasChem.jl test/test_oh_production.jl:181-224, third iteration with rh=50). H₂O = 0.5 · 0.031 · 2.5e25 = 3.875e23 m⁻³ (50% RH — note this differs from the typical-default H₂O=4.0e23 by 3% because the typical-default uses a slightly different ξ_H₂O_sat). Book value ε_OH ≈ 0.23 at rtol=0.1; this test pins the exact computed 0.21252726706138983 from Eq. 6.4 at rel=1e-10 (within rtol=0.1 of the book value — relative error 7.6%). This is the upstream test's tightest case — the book and the exact-formula values are closest at 50% RH where neither limit dominates the denominator.",
      "parameter_overrides": {
        "H2O": 3.875e+23
      },
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "ε_OH",
          "time": 0.0,
          "expected": 0.21252726706138983
        }
      ]
    },
    {
      "id": "book_table_rh80_p207_extension",
      "description": "Extends the upstream book-table coverage to the high-RH 80% case that the upstream test (test/test_oh_production.jl:211) deliberately omits noting `Allow 10% tolerance for k3_eff vs k3 difference and exact vs approx formula. We test at low/moderate RH where exact and approximate forms agree well.` At RH=80%, H₂O = 0.8 · 0.031 · 2.5e25 = 6.2e23 m⁻³. Book value 0.38; computed via exact Eq. 6.4 ε_OH = 0.3196625263651278 (relative error -16% from the book — outside the upstream rtol=0.1 bound, which is why the upstream test omits this case). This .esm test pins the exact computed value at rel=1e-10 to characterize the high-RH divergence — useful regression target for any future updates that change k₃_N₂/k₃_O₂/k₄ defaults or rebalance the k₃_eff weighting. Documents the limit of the book's approximate formula vs the exact one.",
      "parameter_overrides": {
        "H2O": 6.2e+23
      },
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "ε_OH",
          "time": 0.0,
          "expected": 0.3196625263651278
        }
      ]
    },
    {
      "id": "upstream_o1d_linear_in_o3_baseline",
      "description": "Baseline endpoint of the upstream `OHProduction: Qualitative - O1D Proportional to O3` test (GasChem.jl test/test_oh_production.jl:226-247). With O₃=1e18, H₂O=4e23, M=2.5e25 (default scenario except O₃ overridden to baseline 1e18), pins O₁D=74534.16149068323 — same as upstream_eq61_o1d_steady_state_defaults but separately listed so it pairs with `upstream_o1d_linear_in_o3_doubled` to reproduce the upstream ratio assertion O₁D(O₃=2e18)/O₁D(O₃=1e18) ≈ 2.0 (upstream rtol=1e-6). By construction Eq. 6.1 is exactly linear in O₃ (only the numerator depends on O₃, the denominator does not), so the ratio is exactly 2 by IEEE-754 multiplication-reproducibility — both endpoints pinned at rel=1e-10 → the ratio reproduction is structural rather than numerical.",
      "parameter_overrides": {
        "O3": 1e+18
      },
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "O1D",
          "time": 0.0,
          "expected": 74534.16149068323
        }
      ]
    },
    {
      "id": "upstream_o1d_linear_in_o3_doubled",
      "description": "Doubled endpoint of the upstream `OHProduction: Qualitative - O1D Proportional to O3` test. With O₃ doubled to 2.0e18 m⁻³ (other inputs unchanged from default), O₁D doubles to exactly 149068.32298136645 m⁻³ — exactly 2× the baseline 74534.16149068323. Pinned at rel=1e-10. Together with `upstream_o1d_linear_in_o3_baseline` reproduces the upstream linearity assertion: O₁D(2·O₃) = 2·O₁D(O₃) follows from the structural fact that Eq. 6.1's numerator is linear in O₃ and the denominator is independent of O₃. ε_OH and P_OH change with O₃ as: ε_OH is independent of O₃ (denominator: k₃_eff·M + k₄·H₂O — same as baseline), P_OH is linear in O₃ (P_OH = j_O₃·O₃·ε_OH doubles).",
      "parameter_overrides": {
        "O3": 2e+18
      },
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "O1D",
          "time": 0.0,
          "expected": 149068.32298136645
        },
        {
          "variable": "ε_OH",
          "time": 0.0,
          "expected": 0.21863354037267077
        },
        {
          "variable": "P_OH",
          "time": 0.0,
          "expected": 26236024844720.49
        }
      ]
    },
    {
      "id": "zero_H2O_eps_OH_collapses",
      "description": "Boundary case H₂O = 0 m⁻³ → ε_OH = 0 / P_OH = 0 exactly (no OH produced — every O(¹D) is quenched). The 2·k₄·H₂O numerator collapses to IEEE-754 exact zero. O₁D = j_O₃·O₃/(k₃_eff·M) = 6e-5·1e18/(2.868e-17·2.5e25) = 83682.00836820084 m⁻³ — the asymptotic dry-air limit. This pinning uses abs=1e-12 tolerance for the zero-valued outputs to gate the stuck-at-non-zero failure mode (a regression where the 2·k₄·H₂O term is accidentally rewritten to use a non-zero constant), and rel=1e-10 for the non-zero O₁D. Physically H₂O=0 is the dry-stratosphere or dust-storm limit — useful sentinel for any future migration that adds a numerator-bias term.",
      "parameter_overrides": {
        "H2O": 0.0
      },
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "O1D",
          "time": 0.0,
          "expected": 83682.00836820084
        },
        {
          "variable": "ε_OH",
          "time": 0.0,
          "expected": 0.0,
          "tolerance": {
            "abs": 1e-12
          }
        },
        {
          "variable": "P_OH",
          "time": 0.0,
          "expected": 0.0,
          "tolerance": {
            "abs": 1e-12
          }
        }
      ]
    },
    {
      "id": "zero_O3_no_OH_production",
      "description": "Boundary case O₃ = 0 m⁻³ → O₁D = 0 / P_OH = 0 exactly (no O(¹D) source — j_O₃·O₃ numerator collapses to zero). ε_OH = 2·k₄·H₂O/(k₃_eff·M + k₄·H₂O) = 0.21863354037267077 — unchanged from defaults, since ε_OH is independent of O₃ (a structural property of Eq. 6.4). This sentinel gates the stuck-at-non-zero failure mode for O₁D and P_OH; ε_OH must remain at its default value confirming the O₃-independence of the yield. Physically O₃=0 is the pre-photolysis or O₃-titrated limit.",
      "parameter_overrides": {
        "O3": 0.0
      },
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "O1D",
          "time": 0.0,
          "expected": 0.0,
          "tolerance": {
            "abs": 1e-12
          }
        },
        {
          "variable": "ε_OH",
          "time": 0.0,
          "expected": 0.21863354037267077
        },
        {
          "variable": "P_OH",
          "time": 0.0,
          "expected": 0.0,
          "tolerance": {
            "abs": 1e-12
          }
        }
      ]
    },
    {
      "id": "high_pressure_quenching_dominates",
      "description": "High-pressure limit M = 1.0e26 m⁻³ (~4× STP, hypothetical deep-troposphere or surface-pressure-cycle case — Mars's surface during dust storms reaches similar M·CO₂ values, though in pure CO₂ the k₃ values would differ). The k₃_eff·M term grows 4× to 2.868e9 s⁻¹ — far larger than the unchanged k₄·H₂O=8.8e7 s⁻¹, so quenching dominates and ε_OH drops to 0.05954 (about a quarter of the typical-pressure value). O₁D drops 4× to 20297.7 m⁻³ (denominator scales linearly with M); P_OH drops by the same ε_OH·O₁D combined factor to 3.57e12 m⁻³/s. Pinned at rel=1e-10. Demonstrates that increasing M (= increasing pressure / lower altitude) suppresses tropospheric OH production by suppressing the OH yield — the canonical 'why does OH production peak in the mid-troposphere not at the surface' physics.",
      "parameter_overrides": {
        "M": 1e+26
      },
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "O1D",
          "time": 0.0,
          "expected": 20297.699594046007
        },
        {
          "variable": "ε_OH",
          "time": 0.0,
          "expected": 0.05953991880920163
        },
        {
          "variable": "P_OH",
          "time": 0.0,
          "expected": 3572395128552.0977
        }
      ]
    },
    {
      "id": "low_pressure_stratosphere_regime",
      "description": "Stratospheric regime M = 1.0e24 m⁻³ (~25× lower than STP, corresponding to ~30 km altitude) with H₂O = 1.0e21 m⁻³ (typical stratospheric water vapor, 4-5 ppm — far drier than tropospheric). Pins O₁D=2.076e6 m⁻³ (~28× larger than at typical surface because the k₃_eff·M denominator drops by 25× while j_O₃·O₃ stays constant — stratospheric O(¹D) is much more abundant per O₃ molecule), ε_OH=0.01522 (smaller than tropospheric because H₂O is much smaller — only 1.5% of O(¹D) makes OH at these conditions), P_OH=9.13e11 m⁻³/s (an order of magnitude smaller than tropospheric but not negligible). Pinned at rel=1e-10. Anchors the stratospheric-OH limit and exercises a regime where both M and H₂O are simultaneously perturbed — useful sentinel against single-input regression.",
      "parameter_overrides": {
        "M": 1e+24,
        "H2O": 1e+21
      },
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "O1D",
          "time": 0.0,
          "expected": 2076124.5674740488
        },
        {
          "variable": "ε_OH",
          "time": 0.0,
          "expected": 0.015224913494809691
        },
        {
          "variable": "P_OH",
          "time": 0.0,
          "expected": 913494809688.5814
        }
      ]
    },
    {
      "id": "pure_N2_atmosphere_f_O2_zero_sentinel",
      "description": "Hypothetical pure-N₂ atmosphere sentinel: f_N₂=1.0, f_O₂=0.0 → k₃_eff = k₃_N₂ = 2.6e-17 m³/s (no O₂ quenching channel). The denominator shrinks slightly: 2.6e-17·2.5e25 + 2.2e-16·4e23 = 6.5e8 + 8.8e7 = 7.38e8 (8.4% smaller than the well-mixed-atmosphere case). O₁D rises 9% to 81300.81300813008, ε_OH rises 9% to 0.23848238482384823 (because the k₄·H₂O numerator term grows 9% relative to the smaller k₃_eff·M denominator), P_OH rises 9% to 1.43e13 m⁻³/s. Pinned at rel=1e-10. Demonstrates that f_O₂ override behaves correctly as a linear multiplier in k₃_eff (gating the regression where one of f_N₂/f_O₂ is accidentally hard-coded into the equation rather than referenced from the parameter slot).",
      "parameter_overrides": {
        "f_N2": 1.0,
        "f_O2": 0.0
      },
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "assertions": [
        {
          "variable": "O1D",
          "time": 0.0,
          "expected": 81300.81300813008
        },
        {
          "variable": "ε_OH",
          "time": 0.0,
          "expected": 0.23848238482384823
        },
        {
          "variable": "P_OH",
          "time": 0.0,
          "expected": 14308943089430.895
        }
      ]
    }
  ],
  "analyses": [
    {
      "id": "oh_yield_vs_humidity_canonical_curve",
      "description": "OH yield ε_OH and OH production rate P_OH as functions of water vapor concentration H₂O, swept linearly from 1e22 to 1.5e24 m⁻³ (covering ~3% to ~200% of S&P typical 50%-RH at 298 K). Reproduces the upstream doc figure `oh_yield_humidity.svg` (docs/src/oh_production.md, lines 73-122) at the upstream's specific conditions O₃ = 50 ppb at STP M = 2.5e25 m⁻³ → O₃ = 50e-9 · 2.5e25 = 1.25e18 m⁻³. The ε_OH curve rises monotonically from near-zero (deep linear-numerator regime, k₄·H₂O << k₃_eff·M) through ~0.2 at typical 50%-RH conditions and toward the high-RH saturation at ε_OH ≈ 2 (the k₄·H₂O numerator catches up to k₃_eff·M then dominates — every O(¹D) makes 2 OH). The P_OH curve has the same shape, scaled by the constant j_O₃·O₃ factor.",
      "parameters": {
        "O3": 1.25e+18
      },
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "parameter_sweep": {
        "type": "cartesian",
        "dimensions": [
          {
            "parameter": "H2O",
            "range": {
              "start": 1e+22,
              "stop": 1.5e+24,
              "count": 60,
              "scale": "linear"
            }
          }
        ]
      },
      "plots": [
        {
          "id": "eps_OH_vs_H2O",
          "type": "line",
          "description": "OH yield ε_OH vs water vapor number density H₂O at O₃=50 ppb, M=2.5e25 (S&P p. 207 conditions). Monotonically rising curve from ~0.01 at H₂O=1e22 (dry-air limit) through ~0.2 at H₂O=5e23 (50% RH) toward ~0.5 at H₂O=1.5e24 (very wet — the k₄·H₂O term becomes comparable to k₃_eff·M).",
          "x": {
            "variable": "H2O",
            "label": "Water vapor concentration H₂O (m⁻³)"
          },
          "y": {
            "variable": "ε_OH",
            "label": "OH yield ε_OH (dimensionless)"
          }
        },
        {
          "id": "P_OH_vs_H2O",
          "type": "line",
          "description": "OH production rate P_OH vs water vapor number density H₂O at O₃=50 ppb, M=2.5e25 (S&P p. 207 conditions). Same shape as ε_OH (since P_OH = j_O₃·O₃·ε_OH and the j_O₃·O₃ factor is constant in this sweep), scaled by 6e-5·1.25e18 = 7.5e13 → P_OH ranges from ~1e12 at the dry limit to ~4e13 at the high-humidity limit (in m⁻³·s⁻¹).",
          "x": {
            "variable": "H2O",
            "label": "Water vapor concentration H₂O (m⁻³)"
          },
          "y": {
            "variable": "P_OH",
            "label": "OH production rate P_OH (m⁻³·s⁻¹)"
          }
        }
      ]
    },
    {
      "id": "eps_OH_vs_relative_humidity_298K",
      "description": "OH yield ε_OH as a function of relative humidity at 298 K, swept from 5% to 95% RH using the 298 K saturation H₂O mixing ratio ξ_H₂O_sat ≈ 0.031. H₂O = RH · ξ_H₂O_sat · M, with M=2.5e25. Reproduces the S&P p. 207 Table {(10%, 0.047), (25%, 0.12), (50%, 0.23), (80%, 0.38)} at finer RH resolution, surfacing the rising-curve shape that interpolates between the book's discrete points. Useful comparison with the upstream doc page's `oh_production.md` Table reproducing the same data.",
      "parameters": {
        "O3": 1e+18,
        "M": 2.5e+25
      },
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "parameter_sweep": {
        "type": "cartesian",
        "dimensions": [
          {
            "parameter": "H2O",
            "range": {
              "start": 3.875e+22,
              "stop": 7.3625e+23,
              "count": 50,
              "scale": "linear"
            }
          }
        ]
      },
      "plots": [
        {
          "id": "eps_OH_vs_H2O_at_298K",
          "type": "line",
          "description": "OH yield ε_OH vs water vapor H₂O at 298 K equivalent RH (5%-95%) with O₃=1e18 m⁻³, M=2.5e25 m⁻³. The book's S&P p. 207 anchor points are at H₂O = {7.75e22, 1.94e23, 3.88e23, 6.2e23} m⁻³ corresponding to RH = {10%, 25%, 50%, 80%}.",
          "x": {
            "variable": "H2O",
            "label": "H₂O number density (m⁻³, corresponds to RH at 298 K)"
          },
          "y": {
            "variable": "ε_OH",
            "label": "OH yield ε_OH (dimensionless)"
          }
        }
      ]
    },
    {
      "id": "p_oh_vs_o3_linear_scaling",
      "description": "OH production rate P_OH as a function of ozone O₃, swept linearly from 1e17 to 5e18 m⁻³ (covering ~4 ppb to ~200 ppb at STP, the tropospheric O₃ range from clean remote to polluted urban). At fixed H₂O=4e23 and M=2.5e25 (typical 50% RH at 298 K), Eq. 6.3 reduces to P_OH = j_O₃·ε_OH·O₃ where ε_OH=0.2186 is constant → exact linear relationship. The resulting plot is a straight line through the origin with slope 6e-5·0.2186 = 1.31e-5 m⁻³·s⁻¹ per m⁻³ O₃, anchoring the O₃-linear scaling of OH production.",
      "parameters": {
        "H2O": 4e+23,
        "M": 2.5e+25
      },
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "parameter_sweep": {
        "type": "cartesian",
        "dimensions": [
          {
            "parameter": "O3",
            "range": {
              "start": 1e+17,
              "stop": 5e+18,
              "count": 50,
              "scale": "linear"
            }
          }
        ]
      },
      "plots": [
        {
          "id": "P_OH_vs_O3_line",
          "type": "line",
          "description": "OH production rate P_OH vs O₃ number density at typical surface conditions. Straight line through origin with slope ~1.31e-5 s⁻¹, range P_OH ∈ [1.3e12, 6.6e13] m⁻³·s⁻¹.",
          "x": {
            "variable": "O3",
            "label": "Ozone O₃ (m⁻³)"
          },
          "y": {
            "variable": "P_OH",
            "label": "OH production rate P_OH (m⁻³·s⁻¹)"
          }
        }
      ]
    },
    {
      "id": "eps_OH_heatmap_humidity_pressure",
      "description": "OH yield ε_OH as a function of both H₂O and M (joint 2D log-log sweep), surfacing the bivariate dependence of the yield on water vapor (numerator) and air density (denominator). H₂O ∈ [1e21, 1e24] m⁻³ (dry stratosphere to wet tropics, 3 decades) × M ∈ [1e24, 1e26] m⁻³ (~30 km altitude to ~surface-overpressure, 2 decades). Surfaces the canonical regime structure: top-left (high H₂O, low M) gives ε_OH → 2 (all O(¹D) reacts with H₂O), bottom-right (low H₂O, high M) gives ε_OH → 0 (all O(¹D) quenched). The diagonal where k₄·H₂O ≈ k₃_eff·M (≈ H₂O/M ≈ 7.65e-9) is the ε_OH ≈ 1 isoline. Useful for visualizing the full physical-conditions space and identifying the realistic-atmospheric regions (mid-troposphere H₂O/M ≈ 0.02, ε_OH ≈ 0.22).",
      "parameters": {
        "O3": 1e+18
      },
      "time_span": {
        "start": 0.0,
        "end": 1.0
      },
      "parameter_sweep": {
        "type": "cartesian",
        "dimensions": [
          {
            "parameter": "H2O",
            "range": {
              "start": 1e+21,
              "stop": 1e+24,
              "count": 25,
              "scale": "log"
            }
          },
          {
            "parameter": "M",
            "range": {
              "start": 1e+24,
              "stop": 1e+26,
              "count": 20,
              "scale": "log"
            }
          }
        ]
      },
      "plots": [
        {
          "id": "eps_OH_heatmap_H2O_M",
          "type": "heatmap",
          "description": "ε_OH as a function of (H₂O, M) on log-log axes. Top-left corner (high H₂O, low M) saturates near ε_OH=2; bottom-right (low H₂O, high M) goes to ε_OH≈0. The realistic surface-troposphere conditions (H₂O≈4e23, M≈2.5e25) sit in the lower-middle-right region where ε_OH ≈ 0.22.",
          "x": {
            "variable": "H2O",
            "label": "Water vapor H₂O (m⁻³, log)"
          },
          "y": {
            "variable": "M",
            "label": "Air number density M (m⁻³, log)"
          },
          "value": {
            "variable": "ε_OH",
            "reduce": "final"
          }
        }
      ]
    }
  ]
}