BIOL3216 / INTERACTIVE PHYSIOLOGY

Too Hot & Too Cold V5

Thermoregulation

How do physiological responses change the heat budget over time?

S = M − C − R − K − E − Qres

Exchange is positive outward · S is positive when heat is stored

Choose an exposure, then predict the response

Every preset selects Rest.

Ready to run

Each run starts at core 36.8°C · skin 33.7°C

Standardized exposure · 60 minutes

No exposure run yet.
Adjust the conditions and press Run exposure.

Prepare next exposure

Indoor
22°C
−20°C+45°C
40%
10%100%
0.1 m/s
Still · 04 m/s

Changing a control prepares the next run. Every exposure resets to the same initial state.

Same person · Fixed clothing

Same person. Same light clothing.

70 kg · 1.8 m² · clothing fixed at 0.5 clo
Core and skin temperatures evolve.
Illustrative surroundingsRun an exposure to see rates
Open a larger view of the displayed scientific figure.
The same standardized adult in changing environments The approved raster-alpha activity figure wears a short-sleeved shirt, full-length trousers and shoes. The separate equipment illustration does not alter the model. Separate environmental layers depict an indoor room, temperate park, arid landscape, lush vegetation, bare cold landscape and snow-covered ground. Heat exchange with the environmentArrows point outward for heat loss and inward for heat gain. Width indicates magnitude on a compressed scale. Evaporation includes regulatory sweat evaporation and skin diffusion. Respiratory exchange is listed separately. All rates are listed in the heat budget below. Radiation R56 WOutward · heat loss Convection C40 WOutward · heat loss Evaporation E19 WOutward · heat lossSweat + diffusion Conduction K≈0 WContact ≈ 0 METABOLIC HEATM = 115 WActivity + shivering
Outward = heat lossInward = heat gainThicker arrow = larger rate*

Illustrative surroundings · no direct sun · fixed clothing insulation.

Physiological effort & outcome

Values appear after a completed exposure.

Predict, then run

S = — W

Observe how temperatures and physiological responses develop during the same 60-minute exposure.

Modelled teaching responses—not individual predictions.

—
No run

Before the wettedness limit

—
No run

Shared 0–1,000 W scale · total skin evaporation includes diffusion.

Sweat secretion
—
Evaporated sweat
—
Cooling from sweat
—
Skin diffusion
—
Shivering heat
—

Secretion is the controller’s output. Only evaporated sweat cools the skin. The model also limits skin wettedness; fully wet potential is not automatically achieved.

Teaching model · all responses are modelled. Evidence depends on the conditions and output; see model notes.

Heat budget

Whole-body W · positive exchange = outward · positive S stores heat
Activity heat
—Prescribed production
Shivering heat
—Modelled production
Total M
—Activity + shivering
Convection C
—Surface exchange
Radiation R
—Surface exchange
Contact K
—Approximated as zero
Skin evaporation E
—Sweat + diffusion
Breathing heat exchange Qres
—Sensible + latent
Whole-body net heat-storage rate S
—Instantaneous rate

Temperature history

Same starting state · same 60-minute exposure

Current run · solid

Run an exposure to follow core and mean skin temperatures.

Core and skin use separate, labelled temperature scales. Compared runs share the scale within each panel.

Inspect figure — Heat Exchange

Larger view of the displayed run. Opening or closing this view does not change the experiment.

Read figure values

Wind chill is context, not air temperature

Wind chill estimates how cold exposed skin feels because wind increases heat loss. It is not the actual air temperature and is not used as an input to this heat-balance calculation.

V4 converts the completed exposure’s air movement from m/s to km/h for the Canadian Wind Chill Index. The index is shown at or below 0°C when wind is greater than zero; a low-wind correction is used below 5 km/h. In calm air it is not calculated.

The official index uses wind measured at 10 m. This classroom illustration uses the same slider speed for both calculations without a height correction. It is an approximate comparison, not a meteorological or exposure prediction. The index is displayed without a degree symbol.

Model assumptions, equations & limitations

V5 · Final release: Bounded visual integration from instructor-approved V4.2; scientific model 4.0 unchanged. Instructor-approved final release · 10 October 2026. The dated interface reviews below are historical.

Interface release · 1 October 2026: V4.1-RC3 was subsequently approved by the instructor as the locked Thermoregulation V4.1 release; scientific model 4.0 is unchanged. The following RC1 review is historical: its graph-label issue was repaired in RC2, and its pending candidate approval no longer describes the current release. Target-context checks not performed remain NOT TESTED.

Interface release-candidate review · 1 October 2026: V4.1-RC1 retains Gate 1B placement and adds Gate 1C figure inspection, annotation-fit and graph-readability changes. Scientific model 4.0 is preserved. Linux Chromium checks passed for source identity, exact Gate 1B numerical regression and inspection state/keyboard behaviour. The 320 px graph has overlapping 45 and 60 min labels under combined user text spacing; a label-position scope decision is required. Native 200%/400% browser zoom, target Mac/phone/projector, spoken screen readers and target-host checks remain outstanding. See the Gate 1C audit for evidence and limits. This candidate awaits instructor review.

Scientific review status · 1 October 2026: Checkpoints A–C reviewed this V4 scientific model. The independent Checkpoint B audit identified no governing-equation, unit-conversion, integration, or scientific-state error requiring an engine change. Gate 1A changes presentation wording and provenance only; the scientific engine remains model version 4.0. Historical development checks retain their original dates and evidence scope.

V4 follows an illustrative adult through a standardized 60-minute air exposure. It couples a core compartment with a mean skin compartment. Temperature signals drive modelled vasomotion, shivering and sweating; environmental exchange then changes those temperatures. A small change in core temperature can coexist with substantial physiological effort. Heat-budget balance does not establish comfort or sustainable thermoregulation.

1. Model identity, evidence and scope

Framework: Gagge/Pierce two-node thermoregulation, adapted from pythermalcomfort 4.6.0, two_nodes_gagge.py (readable implementation documentation). The exact package version is pinned in verification. The documentation’s “stable” link may change. The code’s coefficient set is the reference for this adaptation; it is not identical to every publication called “Gagge 1986”.

Foundational publication: Gagge, Fobelets & Berglund (1986), A standard predictive index of human response to the thermal environment, ASHRAE Transactions 92(2B), 709–731. The original full text was not available for this build; equations were inspected in the implementation and the primary evaluations linked below. This application does not calculate SET, PMV, comfort or safety categories.

Exploration range: air −20 to 45°C, RH 10–100%, air movement 0–4 m/s, and the four activity settings. This is a software input range, not a validated physiological domain. Subzero exposures are explicitly exploratory. High activity, wind and humidity also extend beyond the selected validation checks. Modelled core temperature is a lumped compartment, not specifically rectal, oesophageal or brain temperature.

Takada, Sakiyama & Matsushita (2011), Tables 1–4 and §3 evaluated sedentary steady-state mean skin temperature; their evidence does not directly validate transient core temperature, shivering or blood flow. Doherty & Arens (1988), model evaluations and Appendix A examined core/skin temperatures and wettedness, finding exercise limitations. Neither establishes accuracy of this adapted model in severe cold.

2. Exposure protocol and variables

Each run begins at core 36.8°C and mean skin 33.7°C, the reference implementation’s initial temperatures. These are a standardized model state, not evidence of equilibrium at 22°C or at the chosen preset. There is no hidden acclimation period. All response variables are recalculated from the initial temperatures. No residual temperature, hydration or response history carries over. The complete exposure is exactly 3,600 simulated seconds.

KindVariables and units
Student inputsAir temperature (°C), relative humidity (%), body-level air movement (m/s), prescribed activity metabolism (W).
Fixed teaching choices70 kg, 1.8 m², 0.5 clo, standing exchange geometry, 101,325 Pa. Mean radiant temperature equals air temperature. No direct sun. External mechanical work = 0.
Evolving statesCore and mean skin temperatures (°C).
Algebraic response variablesModel skin blood flow B (L·m⁻²·h⁻¹), effective skin mass fraction α, shivering (W), sweat secretion (g/h), skin wettedness (dimensionless).
Derived exchangeClothing-surface temperature (°C); core–skin transfer, convection, radiation, skin evaporation, respiration and net storage (whole-body W).

Rest/Light/Moderate/Heavy prescribe 115/230/400/600 W respectively. The labels are illustrative, not universal activity definitions. The conversion is met = W / (1.8 × 58.2), approximately 1.10/2.20/3.82/5.73 met. No external work is subtracted. Shivering is additional. Respiration retains the reference implementation’s activity-only metabolic dependence, and does not increase with shivering. Convection includes the reference’s activity-dependent coefficient.

Changing controls prepares the next exposure; it does not alter displayed results. A reference run is a snapshot of its inputs and history. All endpoint values refer to 60 minutes; the graph shows temperatures throughout. S in watts is an instantaneous rate, not accumulated heat or exposure-averaged storage. Comparisons share scales within each temperature panel. Core and skin panels have separate labelled scales.

3. Heat balances, exchange equations and signs

The following equations use fluxes per square metre. The interface multiplies fluxes by area A = 1.8 m² to show whole-body watts. Exchange is positive outward; negative convection, radiation or respiration is inward heat gain. Internal transfer H is positive from core to skin and cancels from the whole-body budget.

m = m_activity + m_shivering
h_core→skin = (5.28 + 1.163 B)(T_core − T_skin)
s_core = m − h_core→skin − q_res
s_skin = h_core→skin − c − r − e_skin
S = A(s_core + s_skin) = M − C − R − K − E − Q_res

dT_core/dt = A s_core / [c_body mass (1 − α)]
dT_skin/dt = A s_skin / [c_body mass α]
c_body = 0.97 Wh/(kg·K) × 3600 = 3492 J/(kg·K)
α = 0.0417737 + 0.7451833 / (B + 0.585417)

These are effective compartment capacities. Because α changes with blood flow, total stored energy is obtained by integrating S, not by treating mass × c_body × [α T_skin + (1−α) T_core] as a fixed-compartment heat content. Internal tissue conductance is included in H. External contact conduction K is approximated as zero; conduction through clothing is already represented by clothing resistance and is not added again.

v_eff = max(v, 0.1)                     [m/s]
h_c = max(3, 8.600001 v_eff^0.53,
          5.66 (met − 0.85)^0.39)        [W/(m²·K); activity term if met > 0.85]
f_cl = 1 + 0.15 clo;  R_cl = 0.155 clo   [m²·K/W]
h_r = 4 × 0.95 × σ × 0.73 × [(T_cl+T_r)/2 + 273.15]^3
R_a = 1 / [f_cl(h_c + h_r)]
T_cl = (R_a T_skin + R_cl T_air)/(R_a + R_cl)  [T_r = T_air]
c = f_cl h_c(T_cl − T_air)
r = f_cl h_r(T_cl − T_r)

q_res = 0.0023 m_activity(44 − p_air)
      + 0.0014 m_activity(34 − T_air)     [latent + sensible; p in Torr]

The clothing equation is iterated to a 10⁻⁹°C numerical tolerance. The radiation coefficient is the reference’s linearized relation, evaluated at the current clothing and radiant temperatures. It is not V3.2’s quartic implementation. A zero-wind control value still uses 0.1 m/s in physiological exchange, as documented by the reference model. Wind chill uses the supplied speed separately.

Sources: reference kernel: heat-transfer and compartment-update sections; Doherty & Arens, equations 14–35.

4. Vasomotion, shivering, sweat and evaporation

Positive warm/cold signals are departures from the reference core/skin temperatures; negative signals are replaced with zero. These are empirical model response rules. They do not simulate neural pathways, sweat-gland kinetics or all mechanisms of cold defence.

warm_core = max(0, T_core − 36.8); cold_core = max(0, 36.8 − T_core)
warm_skin = max(0, T_skin − 33.7); cold_skin = max(0, 33.7 − T_skin)
B = bound[(6.3 + 120 warm_core)/(1 + 0.5 cold_skin), 0.5, 90]
T_body = α T_skin + (1 − α) T_core
sweat_secretion = min[500, 170 max(0,T_body−36.49) exp(warm_skin/10.7)]
                                                        [g/(m²·h)]
m_shivering = 19.4 cold_skin cold_core                    [W/m²]

B represents model skin blood flow, with L/(m²·h) supported by the dimensional relation using blood heat capacity 1.163 Wh/(L·K). It is not a measured limb flow or a percentage of vasodilation. Sweat units follow 0.68 Wh/g latent heat; the reference documentation’s kg/h/m² label is inconsistent with this conversion. V4 displays g/h for the whole body. The blood and secretion bounds are model parameters, not universal physiological maxima.

Shivering requires both cold signals in this formulation. It has no V3.2 460 W ceiling, no fatigue or exhaustion process and no cold-skin-rate signal. This is a material limitation of cold-response interpretation. Activity may suppress model shivering through its temperature effects, rather than by subtracting activity from a fixed metabolic ceiling.

p_sat(T) = exp[18.6686 − 4030.183/(T+235)]           [Torr]
p_air = RH/100 × p_sat(T_air)
R_e = 1/(2.2 f_cl h_c) + R_cl/(2.2 × 0.45)
e_max = [p_sat(T_skin) − p_air]/R_e                 [W/m²]
w_max = 0.59 v_eff^(−0.08)                         [clothed reference]
e_sweat = min[0.68 × sweat_secretion,
              e_max × max(0,(w_max−0.06)/0.94)]     [for e_max > 0]
e_diffusion = 0.06(e_max − e_sweat)
e_skin = e_sweat + e_diffusion
w = 0.06 + 0.94(e_sweat/e_max)
evaporated_sweat = e_sweat/0.68                    [g/(m²·h)]

Secretion and evaporation remain separate. Fully wet environmental capacity e_max is a potential; the model’s maximum wettedness usually imposes a lower limit. Diffusion is obligatory skin water loss, separate from regulatory sweat. The difference between secreted and evaporated sweat is not a prediction of clothing wetness, retained liquid or runoff. No dehydration history is calculated.

Documented correction: the reference’s saturated branch uses p_rsw = w_max/0.94, inconsistent with its own w = 0.06 + 0.94 p_rsw. V4 uses (w_max−0.06)/0.94 and never evaporates more regulatory sweat than is secreted. At zero capacity evaporation is exactly zero, without a small positive substitute. Negative capacity stops the exposure: condensation heat is outside the represented processes.

Sources: reference response and evaporation branches; Doherty & Arens, equations 25–30 for sweat/diffusion bookkeeping. The correction is algebraic, not a new fitted coefficient.

5. Parameter provenance and adaptations
Literature/implementation parameters are identified separately from teaching and numerical choices. R denotes the pinned pythermalcomfort 4.6.0 kernel; D denotes Doherty & Arens (1988). The code retains the exact values below.
ParameterValue and unitSource location / rationale
Mass / area / clothing70 kg / 1.8 m² / 0.5 cloBIOL3216 teaching choices inherited from V3.2. Area overrides R’s default.
Pressure / geometry / work101325 Pa / standing / 0 WFixed exposure choices; R’s standing radiation-area factor.
Initial and signal temperaturesCore 36.8°C; skin 33.7°C; mean-body signal reference 36.49°CR initialization; body reference uses 0.1 × 33.7 + 0.9 × 36.8.
Body specific heat0.97 Wh/(kg·K)R compartment capacities; D equations 33–35. Converted to J/(kg·K) for seconds.
Tissue / blood heat transfer5.28 W/(m²·K); 1.163 Wh/(L·K)R core–skin transfer; D equation 30.
Blood-flow ruleBasal 6.3; bounds 0.5–90 L/(m²·h); dilation 120 L/(m²·h·K); constriction 0.5 K⁻¹R control coefficients. Other Gagge versions differ; they were not substituted.
Effective skin fraction0.0417737 + 0.7451833/(B+0.585417)R alpha update; units of constants follow B so α is dimensionless.
Sweating / shivering170 g/(m²·h·K); exponent scale 10.7 K; secretion cap 500 g/(m²·h); shivering gain 19.4 W/(m²·K²)R thermal-state controller; empirical response parameters.
EvaporationLatent heat 0.68 Wh/g; Lewis ratio 2.2 K/Torr at sea level; clothing permeability 0.45; diffusion 0.06; wettedness 0.59 v_eff⁻⁰·⁰⁸R evaporation section; wind in m/s. Complementary sweat fraction 0.94 = 1−0.06.
Vapour pressure18.6686, 4030.183, 235 in the stated Torr expressionR saturation formula; V4 replaces V3.2’s Buck expression to retain one framework. Liquid-water reference, including below zero; subzero use is exploratory.
Dry exchangeσ = 5.6697×10⁻⁸ W/(m²·K⁴); emissivity 0.95; radiation area 0.73; R_cl = 0.155 clo; f_cl = 1+0.15 cloR clothing/heat-transfer section. All convection constants appear in §3.
Respiration0.0023 Torr⁻¹ × (44 Torr−p_air); 0.0014 K⁻¹ × (34°C−T_air)R respiration equations, multiplied by activity metabolism.
Exposure / integrator3600 s / fourth-order Runge–Kutta, 5 s steps / history every 60 sTeaching and numerical choices. Smaller-step checks are reported in §6.
Display conventionsTemperatures 2 decimals; rates rounded W; near-zero S if |S| < 0.5 WReadability choices, not physiological thresholds. Computation uses unrounded values.

Adaptations beyond the evaporation correction: controller variables and effective capacities are evaluated consistently at each integration stage. This removes the reference’s one-step lag and arbitrary initial evaporation. At t = 0 the temperatures are unchanged, B = 6.3 and α is derived from B (about 0.150), rather than held at the reference’s first-step 0.1. The mean-body signal reference remains 36.49°C. Endpoint rates are recomputed at the endpoint temperature state.

The reference loop labelled 60 minutes performs 59 one-minute updates; V4 integrates exactly 3600 s and records t = 0. Its smaller-step solver and tighter clothing iteration are numerical adaptations. These differences are isolated in the comparison checks; numerical equality with the original wrapper is not claimed. The app retains no SET/comfort-index calculations or V3.2 compensation rules.

6. Evidence, verification and outstanding limitations

Build checks, 24 September 2026. These results belong to this V4 adaptation, rather than V3.2.

  • Mathematical verification: 1,200 combinations spanning 12 temperatures, five humidities, five wind speeds and four activity levels. 1,080 exposures completed; 120 were explicitly stopped because condensation is not represented; no numerical failures. Maximum instantaneous whole-body balance residual was 3.07e-13 W. Sweat evaporation never exceeded secretion, diffusion remained nonnegative, wettedness limits held, and blood flow remained within the model bounds.
  • Reference implementation: a test-only, unmodified-loop JavaScript port reproduced 12 pythermalcomfort 4.6.0 Python cases across 12 physiological output fields, with maximum absolute difference 2.56e-12 in the native units. This port is only a verification baseline; the shipped V4 includes the documented adaptations.
  • Numerical sensitivity: 10 s, 5 s and 2.5 s integration steps were compared in 12 representative exposures. Halving the shipped 5 s step changed endpoint temperatures by less than 0.000001°C, heat rates by less than 0.001 W, and secretion by less than 0.001 g/h. The largest temperature difference across the recorded histories was less than 0.00002°C. This is numerical convergence, not physiological accuracy.
  • Human-data check: all 56 condition means in Takada et al. Table 1 were compared with the adapted controller. Mean skin-temperature RMSE was 1.12°C, signed bias -0.55°C (model minus observed), and largest absolute difference 2.77°C. No coefficients were fitted. These errors are material and are not individual prediction intervals.

The human-data harness followed Table 3’s comparison assumptions: 81.7 kg, 1.96 m², h_c = 3.1 and h_r = 4.65 W/(m²·K), plus each condition’s clothing and metabolism. It ran for 10 hours to approximate the reported steady-state comparison. The largest terminal skin-temperature drift was 0.014°C/h. These overrides are confined to verification; the student app remains 70 kg, 1.8 m², 0.5 clo and 60 minutes. Several original metabolic rates were estimated, some environmental details were assumed, and the observations are rounded group means. This is a limited reproduction check using a published compilation, not participant-level or independent transient validation.

Outstanding physiological checks: no direct transient human-data validation was completed for this build’s core temperature, sweat secretion, blood flow or shivering. No severe-cold validation was completed. The lower error reported for a different 1986 coefficient set in the paper is not claimed for V4. The 60-minute protocol and changed numerical implementation still need appropriate transient datasets before stronger predictive claims are made.

Application verification: Chromium desktop/laptop, tablet and 390/320 px phone layouts were rendered and inspected. Automated checks covered pending inputs, unchanged displayed artwork, deterministic resets, immutable comparison snapshots, replace/remove actions, keyboard tabs/sliders, dialog closing/focus return, and condensation messaging. No script errors or application network requests occurred during the offline test. Automated WCAG A/AA checks found 0 violations in the tested main view and 0 in the expanded reference dialog. Wide tables are keyboard-focusable. This does not replace a full assistive-technology or cross-browser audit.

See preset results and the effect of adaptations
All presets at Rest. Original reference: 59 one-minute updates, including its evaporation branch. V4: corrected evaporation and exactly 60 minutes. This table reports differences; it is not a requirement that the models agree.
PresetReference core °CV4 core °CReference skin °CV4 skin °CV4 skin evaporation W
indoor36.83936.83932.47432.47114.97
hot-dry36.96536.96535.33035.329112.91
hot-humid37.20337.24936.41336.55285.68
warm-windy36.85936.85934.02734.02747.79
cold-still36.65736.65724.25924.29511.55
cold-windy36.61236.61219.34519.42017.25

The hot/humid preset illustrates the evaporation correction. Its core/skin outputs progress as follows when changes are isolated:

Calculation stageCore °CSkin °C
Unmodified reference37.203436.4127
Correct evaporation; still 59 min37.249136.5521
Correct evaporation; 60 one-minute updates37.253936.5613
Reference order; 720 five-second updates37.249436.5528
V4 coherent controller and RK4; 60 min37.249036.5520

Different intermediate results reflect the documented timing, numerical and bookkeeping changes. All response gains remain pinned.

Inspect the 56 human-data comparison rows
Takada et al. Table 1 group means. Skin temperatures are °C; difference = V4 harness minus observed. Met is converted with 58.2 W/m². The adapted controller and exchange definitions remain V4’s.
StudyAir °CRH %clo / metObserved skinModel skinDifference
Authors29.4470.06 / 134.334.10-0.20
Authors26.346.10.06 / 133.432.31-1.09
Authors38.868.10.06 / 136.137.85+1.75
Mori 200326.8470.06 / 131.532.80+1.30
Mori 200331.149.60.06 / 134.534.50+0.00
Mori 200330.443.60.06 / 134.534.31-0.19
Tamura 198422500.1 / 0.729.429.00-0.40
Tamura 198425500.1 / 0.730.930.75-0.15
Tamura 198428500.1 / 0.733.032.59-0.41
Tamura 198431500.1 / 0.734.234.41+0.21
Tamura 198434500.1 / 0.734.434.99+0.59
Minami 200835500.5 / 135.535.42-0.08
Minami 200835700.5 / 135.736.12+0.42
Minami 200840500.5 / 136.137.51+1.41
Ogawa 197421.8600.06 / 0.931.628.83-2.77
Ogawa 197424.2600.06 / 0.932.630.28-2.32
Ogawa 197426.9600.06 / 0.933.332.45-0.85
Ogawa 197429.2600.06 / 0.934.134.06-0.04
Ogawa 197421.8600.1 / 0.830.628.98-1.62
Ogawa 197424.2600.1 / 0.832.030.39-1.61
Ogawa 197426.9600.1 / 0.832.832.08-0.72
Ogawa 197429.2600.1 / 0.833.934.09+0.19
Ogawa 197421.8600.6 / 0.832.830.85-1.95
Ogawa 197424.2600.6 / 0.833.432.15-1.25
Ogawa 197426.9600.6 / 0.834.234.15-0.05
Ogawa 197429.2600.6 / 0.834.834.49-0.31
Ogawa 197521.8600.06 / 0.930.528.83-1.67
Ogawa 197524.2600.06 / 0.931.730.28-1.42
Ogawa 197526.9600.06 / 0.932.732.45-0.25
Ogawa 197529.2600.06 / 0.933.534.06+0.56
Ogawa 197521.8600.1 / 0.830.328.98-1.32
Ogawa 197524.2600.1 / 0.831.730.39-1.31
Ogawa 197526.9600.1 / 0.832.632.08-0.52
Ogawa 197529.2600.1 / 0.834.034.09+0.09
Ogawa 197521.8600.6 / 0.831.630.85-0.75
Ogawa 197524.2600.6 / 0.832.532.15-0.35
Ogawa 197526.9600.6 / 0.833.234.15+0.95
Ogawa 197529.2600.6 / 0.834.034.49+0.49
Ogawa 197619.5600.6 / 0.831.429.69-1.71
Ogawa 197622600.6 / 0.832.130.95-1.15
Ogawa 197624.5600.6 / 0.833.032.33-0.67
Ogawa 197627600.6 / 0.833.934.16+0.26
Ogawa 197717.1600.6 / 0.830.528.52-1.98
Ogawa 197719.6600.6 / 0.831.129.74-1.36
Ogawa 197721.8600.6 / 0.832.030.85-1.15
Ogawa 197724.2600.6 / 0.832.932.15-0.75
Arita 198916.1500.6 / 130.128.06-2.04
Arita 198920.5500.6 / 132.330.31-1.99
Arita 198924.4500.6 / 133.633.79+0.19
Arita 198927.7500.6 / 134.834.33-0.47
Yokoyama 200222720.06 / 130.329.06-1.24
Yokoyama 200224640.06 / 132.030.36-1.64
Yokoyama 200226570.06 / 132.732.20-0.50
Yokoyama 200228510.06 / 133.333.79+0.49
Yokoyama 200230450.06 / 134.034.23+0.23
Yokoyama 200232400.06 / 134.234.66+0.46

Pinned reference file SHA-256: a925ade548d67d7bdfe738ee563dd864577dc70cb573155e38c49ff66a8e287f. Numerical checks used unrounded values; displayed tables are rounded.

Evidence categories: equation checks establish mathematical consistency; reference comparisons establish implementation agreement or quantified differences; human-data checks assess agreement with particular observations. These are distinct. A numerical benchmark or plausible graph is not human-data validation. A published model’s validation does not automatically transfer to V4’s adaptations.

Unsupported processes include regional temperatures and peripheral tissue injury, direct solar radiation, external contact/immersion, wet clothing, hydration dynamics, acclimatization, individual susceptibility, exposure changes during a run, and fatigue or exhaustion of responses. There are no survival, frostbite, hypothermia-countdown or clinical/safety outputs. Temperature outputs are not silently clamped.

Blood-flow bounds, secretion bounds and the absence of a shivering maximum are model assumptions. The selected human checks do not directly establish the accuracy of these variables. Severe-cold and high-activity runs remain exploratory even when numerical convergence is excellent. There are no individual prediction intervals.

7. Wind chill, illustration and accessibility

The Canadian Wind Chill Index remains separate from the engine. For T ≤ 0°C, wind v > 0 and V = 3.6v km/h:

V ≥ 5: WCI = 13.12 + 0.6215T − 11.37V^0.16 + 0.3965TV^0.16
0 < V < 5: WCI = T + [(−1.59 + 0.1345T)/5]V

Calm air and positive temperatures have no index. The index uses the run’s supplied speed as a stand-in for an official 10 m observation, with no height correction, and is displayed without a degree symbol. It is never used as air temperature in the physiological model. Sources: ECCC glossary and ECCC Wind Chill Index.

The approved V5 activity figures and layered treadmill replace the adult illustration. The V4.2 environmental artwork and embedded fonts are retained. Poses and equipment are illustrative only; clothing remains fixed at 0.5 clo. Scenery does not enter the equations. Snow and vegetation are illustrative, not predicted weather or contact heat. Arrow width is 3 + 11√[min(|W|,600)/600] illustration pixels; it is compressed and capped. The number remains the calculated rate. Near-zero displayed flows use neutral dotted lines. Direction labels supplement blue heat loss and orange heat gain.

Temperature traces use labelled core/skin panels, with solid/current and dashed/reference traces and different endpoint symbols. A temperature table provides the graph’s values. Controls support the keyboard; arrow keys switch the two view tabs. Dialogs close with their Close button or Escape. Reduced-motion settings stop animation. All assets are embedded; only source links need the internet. No student responses, identifiers, telemetry or run histories are collected or stored.

8. Sources and software attribution
  1. pythermalcomfort 4.6.0 source. Tartarini & Schiavon (2020), pythermalcomfort: A Python package for thermal comfort research, SoftwareX 12, 100578. Publication. MIT licence; adapted physiological kernel.
  2. Doherty & Arens (1988), Evaluation of the physiological bases of thermal comfort models. Equations and human-data comparison; Appendix A provides experimental conditions.
  3. Takada, Sakiyama & Matsushita (2011), Validity of the two-node model for predicting steady-state skin temperature. Building and Environment 46, 597–604. Parameter-version comparison and sedentary skin-temperature dataset.
  4. Mou and colleagues (2023), comparative thermoregulation-model study. IBPSA proceedings, DOI 10.26868/25222708.2023.1723. A limited 26→36°C transient skin-temperature comparison; not validation of this build.
  5. Fiala, Lomas & Stohrer (2001) and JOS-3 author repository: alternatives considered during model selection; not sources of borrowed controller coefficients.

IBM Plex Sans and STIX font licences remain embedded in the HTML. The following licence covers the adapted pythermalcomfort code:

MIT License

Copyright (c) 2019 Federico Tartarini

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice (including the next paragraph) shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.