BIOL3216 / INTERACTIVE PHYSIOLOGY
Thermoregulation
How do physiological responses change the heat budget over time?
Exchange is positive outward · S is positive when heat is stored
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.
Changing a control prepares the next run. Every exposure resets to the same initial state.
Same person. Same light clothing.
Context only · air temperature is unchanged.
Illustrative surroundings · no direct sun · fixed clothing insulation.
Values appear after a completed exposure.
Whole-body net heat-storage rate · at 60 min
Observe how temperatures and physiological responses develop during the same 60-minute exposure.
Modelled teaching responses—not individual predictions.
Before the wettedness limit
Shared 0–1,000 W scale · total skin evaporation includes diffusion.
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.
Vasomotion changes internal heat transfer and the model’s effective skin heat capacity. Skin temperature then affects exchange through the clothing. Blood flow is a whole-body model variable, not a clinical measurement.
Teaching model · all responses are modelled. Evidence depends on the conditions and output; see model notes.
Same starting state · same 60-minute exposure
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.
Larger view of the displayed run. Opening or closing this view does not change the experiment.
Prepared changes are not shown.
Figure: current run only. The retained reference remains in the temperature histories.
Scroll or swipe across the figure. With the figure focused, use the arrow keys.
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.
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.
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”.
two_nodes_gagge.py
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.
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.
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.
met = W / (1.8 × 58.2)
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.
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.
mass × c_body × [α T_skin + (1−α) T_core]
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.
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.
p_rsw = w_max/0.94
w = 0.06 + 0.94 p_rsw
(w_max−0.06)/0.94
Sources: reference response and evaporation branches; Doherty & Arens, equations 25–30 for sweat/diffusion bookkeeping. The correction is algebraic, not a new fitted coefficient.
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.
Build checks, 24 September 2026. These results belong to this V4 adaptation, rather than V3.2.
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.
The hot/humid preset illustrates the evaporation correction. Its core/skin outputs progress as follows when changes are isolated:
Different intermediate results reflect the documented timing, numerical and bookkeeping changes. All response gains remain pinned.
Pinned reference file SHA-256: a925ade548d67d7bdfe738ee563dd864577dc70cb573155e38c49ff66a8e287f. Numerical checks used unrounded values; displayed tables are rounded.
a925ade548d67d7bdfe738ee563dd864577dc70cb573155e38c49ff66a8e287f
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.
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.
3 + 11√[min(|W|,600)/600]
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.
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.