Figure 2 · Steady-state Fs / Fo light responses across drought
Smooth lines: saved Julia model trajectories. White-rimmed circles: subpixel-digitized Flexas et al. (2002) Figure 2 observations.
Why only coupled Cytochrome b₆f mechanics reproduce drought fluorescence: an interactive comparison of six steady-state chlorophyll fluorescence and SIF yield formulations in LeafPhotosynthesis.jl, overlaid on the subpixel-digitized light-response trajectories of Flexas et al. (2002, Physiologia Plantarum 114:231–240, Figs. 2, 3C & 5) across six water-stress stages in Vitis vinifera cv. Tempranillo.
Every curve, residual, peak and RMSE value is computed in Julia when the documentation is built; switching precision loads an independent Float64 or Float32 run. Digitized Flexas et al. (2002) points are strictly observational overlays: no parameter was fitted to them. The Johnson & Berry (2021) rate constants are used unchanged, and the detector weighting εF1/εF2 = 2.0 is carried over from Johnson & Berry (2021), who fitted 2.0 (interquartile range 1.6–2.5) for a LI-6800 on Populus fremontii; it is not a calibration of the Walz PAM-2000 that Flexas et al. used. The following inputs were chosen by the author and have no Tempranillo-specific source in this repository: the non-stomatal Hill schedule (half-conductance 56 mmol H₂O m⁻² s⁻¹, exponent 4, floors 0.60 for Vcmax and 0.85 for Cyt b₆f), Vqmax = 174 μmol e⁻ m⁻² s⁻¹, gm = 0.2 mol CO₂ m⁻² s⁻¹ and Γ* = 3.70 Pa. The schedule matters: with stomatal diffusion alone the lit Figure 2 RMSE is 0.097 and the Inset RMSE 0.144; with the schedule they are 0.062 and 0.038.
Select any fluorescence formulation below to overlay its six drought light-response curves (g = 150, 80, 62, 55, 39, 10 mmol H₂O m⁻² s⁻¹) against the full digitized trajectories from Flexas et al. (2002) Figure 2. Single-variable empirical KN(x) models lock every curve to an identical peak ceiling, whereas two-state Cytochrome b₆f with the PAM observer tracks peak suppression, a leftward peak shift, and high-light sub-Fo quenching simultaneously. Peak positions differ: the model's well-watered peak lies at PPFD = 1050 μmol m⁻² s⁻¹, the digitized one at 729.
Smooth lines: saved Julia model trajectories. White-rimmed circles: subpixel-digitized Flexas et al. (2002) Figure 2 observations.
Choose one stomatal conductance treatment to compare all six fluorescence formulations side by side against that treatment's digitized Figure 2 curve (left), paired with the internal thylakoid redox poise
Simultaneous comparison at the selected drought conductance.
Open PSII traps \(q_2\) govern the low-light rise; Rubisco sink limitation triggers \(K_N > 0\) at the peak.
Beyond the six main light curves, Flexas et al. (2002) tested steady-state fluorescence across continuous stomatal conductance at
Empirical rate-competition models parameterize regulated heat dissipation solely as a function of relative PSII light saturation \(x = 1 - \Phi_P / \Phi_{P0} \in [0, 1]\):
Lowering \(g_{sw}\) or \(V_{c\max}\) rescales where \(x(\mathrm{PPFD})\) lies along the irradiance axis, but every sweep traverses the exact same scalar curve, locking global peak height at \(1.462\) (\(x^* \approx 0.19\), Cotton) or \(1.453\) (\(x^* \approx 0.23\), Drought) at every conductance.
Johnson & Berry (2021) decouple open PSII trap fraction \(q_2(Q_1, V_{q\max})\) from stromal sink yield \(\Phi_{P2}(Q_2, C_c, V_{c\max})\), observed through long-pass (\(>700\;\mathrm{nm}\)) PSI/PSII weights \(\varepsilon_{F1}/\varepsilon_{F2} = 2.0\), carried over from their fit for a LI-6800 on Populus fremontii:
Redox-invariant PSI emission (\(\sim 39\%\) of dark \(F_o\)) yields dark apparent \(F_v / F_m = \mathbf{0.821}\) (measured: \(0.80\text{–}0.83\); the agreement depends on carrying the weighting over to the PAM-2000), while earlier \(A_j \to A_c\) crossover under drought truncates the \(q_2\)-driven rise at progressively lower \(F_s / F_o\).
In ambient air (\(O = 21\;\mathrm{kPa}\)), even at \(C_c = \Gamma_*\) where net carboxylation and photorespiration cancel in carbon uptake, gross oxygenation and cyclic PSI flow (\(\eta(\Gamma_*) = 13/12 \approx 1.083\)) keep consuming electrons:
At \(25\;^\circ\mathrm{C}\), \(J_{P700,c}(\Gamma_*) \approx 0.799\,V_{c\max}\): for \(V_{c\max} = 96\), pure stomatal closure cannot push \(J_{P700,c}\) below \(76.7\;\mu\mathrm{mol\;e^-\;m^{-2}\;s^{-1}}\) (\(\approx 79.3\) as \(g_{sw} \to 0\) with \(R_d = 0.80\)), which is why the model needs a non-stomatal downregulation (an author-chosen schedule) to follow the severe-drought curves (\(g_{sw} < 60\)).
Evaluated across all \(n = 49\) illuminated Figure 2 digitized observations (\(\mathrm{PPFD} > 0\)) and all \(n = 6\) Figure 2 Inset observations (\(\mathrm{PPFD} = 750\;\mu\mathrm{mol\;m^{-2}\;s^{-1}}\)). No parameter was fitted to the digitized curves; the inputs chosen by the author are listed at the top of this page.
Rebuild every table and chart from the repository root; the documentation build runs the benchmark in both precisions:
julia --project=. examples/flexas_2002.jl
julia --project=docs docs/make.jlThe walkthrough gives the complete derivation and API links; the example gallery shows the summary panels next to the other examples.