Plant biophysics uses physical laws and quantitative measurements to explain how plants exchange matter and energy, maintain internal states, and respond to forces. Its most useful question is often deceptively simple: which combination of driving difference, transport pathway, storage and regulation could produce the response we observe? A leaf temperature or a water-flow reading becomes more informative when it is connected to a model that can be tested.
This guide builds that connection through original examples of plant transport, storage, energy exchange and measurement. It concentrates on vascular plants and on models a learner can inspect, rather than attempting a catalogue of every molecular technique. All numerical specimens, parameters and datasets are fictional teaching constructions. They are not species measurements, irrigation recommendations or tree-safety assessments.
THE GOLD STANDARD · PLANT BIOPHYSICS
Connect plant responses with physical explanations
Choose a route through transport, storage, energy and testable models. All teaching chapters remain open below.
ROUTE 1 · CHAPTERS 1–4
Define the physical question
Set boundaries, quantities, balances and transport relationships.
ROUTE 2 · CHAPTERS 5–8
Build and test a model
Compare networks, storage and response times.
ROUTE 3 · CHAPTERS 9–12
Connect physical processes
Link anatomy, energy, coupling and numerical solutions.
ROUTE 4 · CHAPTERS 13–16
Read the evidence
Check uncertainty, comparisons, mechanics and identification.
ROUTE 5 · CHAPTERS 17–20
Diagnose and transfer
Repair explanations, test a changed case and continue learning.
Full chapter index · Try the transfer task · Gold Standard subject hub
Chapter index
Define the physical question · Chapters 1–4
Build and test a model · Chapters 5–8
Connect physical processes · Chapters 9–12
Read the evidence · Chapters 13–16
CHAPTER 1 OF 20 · Define the physical question
1. Start with a system whose boundaries you can name
Imagine a leaf becoming warmer after its surroundings change. You could describe the observation at the scale of a single pore, the whole leaf, a branch or a canopy. Each choice creates a different accounting boundary. Energy crossing a leaf surface is an input or output for a leaf model; an exchange between two parts of a canopy may remain internal to a canopy model.
Begin by writing three statements: what is inside the system, what crosses its boundary, and what can accumulate inside. For a leaf-water model, water entering through its supply pathway and water leaving through evaporation are boundary flows. Changes in stored water belong inside the account. For a mechanical model, external forces and supports matter alongside the geometry of the structure.
The Plant Physiology guide explains the biological processes. Here the additional task is to decide which quantities must be measured together before one explanation can be preferred over another. A model can be small without being careless if its boundary and omissions are explicit.
A useful first drawing contains a box, labelled arrows and one stored quantity. Resist the temptation to label every arrow “energy” or “water movement”. Give it a direction, a quantity and a unit. If the measurement cannot be assigned to the drawing, either the drawing or the proposed interpretation needs revision.
CHAPTER 2 OF 20 · Define the physical question
2. Separate an amount, a rate and a driving difference
Water content, water-flow rate and water-potential difference answer different questions. An amount might be expressed in millilitres, a flow in millilitres per hour, and a potential difference in megapascals. A plant can contain considerable water while transferring it slowly through a particular pathway. Conversely, rapid transfer does not demonstrate a large accessible store.
Build a quantity ledger before calculating. Include the symbol, definition, unit, spatial basis and measurement interval. “Per leaf”, “per square metre of leaf” and “per square metre of ground” cannot be exchanged silently. A conductance calculated for one whole leaf is not automatically comparable with an area-normalised conductance from another experiment.
For an invented comparison, leaf A carries 3 millilitres per hour across 150 square centimetres, while leaf B carries 2 millilitres per hour across 50 square centimetres. Whole-leaf flow is larger in A. Flow per area is 0.020 versus 0.040 millilitres per hour per square centimetre, so B has twice the area-normalised value. Neither statement contradicts the other.
The important question is which quantity answers the investigation. Whole-plant water supply requires an integrated flow. Comparing exchange surfaces may require a surface-based rate. Write the denominator into the conclusion: “higher per unit area” is more informative than “better at transport”. This simple habit prevents many apparently biological disagreements from being disagreements about units.
CHAPTER 3 OF 20 · Define the physical question
3. Use conservation before choosing a complicated mechanism
Conservation provides an initial test that does not require every microscopic process to be known. Over a chosen interval, change in stored water equals inflow minus outflow, provided all relevant water pathways have been included. The statement is an accounting identity; the difficult work is deciding what the measurements actually cover.
Suppose a fictional leaf receives 5 millilitres during an hour and loses 7 millilitres. Its water store falls by 2 millilitres. Equal inflow and outflow would instead imply no net storage change over that hour. It would not prove that storage was constant at every instant: a gain followed by a matching loss could be hidden by the interval average.
Now suppose a researcher measures water entering a cut shoot and calls it instantaneous transpiration. Storage change could separate the two quantities. A leak or an omitted surface could also disturb the comparison. The existing transpiration and plant-water-balance specialist develops the biological pathway and the interpretation of potometer evidence.
A useful repair is to state the balance first, then identify which terms were measured, estimated or assumed negligible. Do not use an unexplained residual as proof of a particular hidden mechanism. A residual tells you the present account does not close. It becomes evidence for a proposed mechanism only when alternative missing terms and measurement errors have been investigated.
CHAPTER 4 OF 20 · Define the physical question
4. Treat conductance as a relationship with conditions
A simple transport model relates flow Q to a driving difference ΔΨ through Q = KΔΨ. Here K is hydraulic conductance for the specified pathway. Equivalently, Q = ΔΨ/R with resistance R = 1/K. These expressions become meaningful only after defining the sign convention, endpoints and units.
In a fictional linear pathway, a difference of 0.8 MPa and conductance of 2 mL per hour per MPa give a flow of 1.6 mL per hour. If the difference changes to 0.4 MPa while conductance stays fixed, flow halves. If flow halves without an endpoint measurement, however, reduced conductance is only one possible explanation. The driving difference may have changed instead.
The word “linear” is a condition, not a compliment. A fixed K means the model uses the same proportionality across the range being considered. A living pathway can change state, so the relation should be tested over the relevant range and duration. A fitted slope from one interval does not establish a permanent property of an entire plant.
For a practical graph-reading exercise, plot flow against measured driving difference. Examine whether one straight line describes the observations and whether the intercept is consistent with the model. Then inspect the order in which measurements were collected. A changing specimen can produce an apparent curve even when the intended comparison assumed a stable pathway.
CHAPTER 5 OF 20 · Build and test a model
5. Discover what endpoint measurements cannot reveal
Consider two resistances in series, with no storage or branching between them. The same steady flow passes through both, and the total driving difference is the sum of their individual drops. Therefore total resistance is R₁ + R₂. This follows directly from writing ΔΨ₁ = QR₁ and ΔΨ₂ = QR₂ and adding the equations.
Let the total difference be 1 MPa, R₁ = 0.25 MPa h/mL and R₂ = 0.75 MPa h/mL. Total resistance is 1, so flow is 1 mL/h. The individual drops are 0.25 and 0.75 MPa. Swap the two resistances and endpoint flow remains exactly the same, but the intermediate potential changes.
This is an original example of an identification problem: endpoint flow and endpoint difference establish the sum, not the allocation. A measurement between the two components could distinguish the alternatives under these assumptions. Naming the components “root” and “shoot” would not supply the missing measurement.
A useful explanation therefore separates what the experiment identifies from what the investigator hopes to learn. If the data only determine total resistance, report total resistance. To localise a change, add an observation capable of separating the possibilities. More decimal places on the same endpoint reading cannot solve a structural lack of information.
CHAPTER 6 OF 20 · Build and test a model
6. Compare changes where they actually occur
Using the same serial network, imagine doubling only R₁ from 0.25 to 0.50 while holding R₂ at 0.75. Total resistance becomes 1.25 and flow under the same 1 MPa difference becomes 0.8 mL/h. Now restore the original network and double only R₂ to 1.50. Total resistance becomes 1.75 and flow becomes about 0.571 mL/h.
Both interventions doubled one component, yet their whole-system consequences differ. The second component originally contributed a larger share of the total resistance. A percentage change without the starting contribution is incomplete information.
Parallel pathways require a different calculation. For two independent paths sharing the same endpoints, flows add and total conductance is K₁ + K₂. If each has conductance 1 mL/h/MPa, their combined conductance is 2. Losing one path halves this simplified total, provided the remaining path and driving difference do not change.
Do not combine series and parallel rules by visual resemblance alone. Trace the actual connections: which components share flow, which share endpoints, and where can material leave or accumulate? A branching leaf network can require more detail than either elementary example. The purpose of these calculations is to make assumptions inspectable, not to turn a complex plant into an electrical circuit without checking the mapping.
A steady-state model can equate supply and loss because it assumes no continuing change in storage. A transient model must allow the two to differ. Introduce stored water W and water potential Ψ, and define a local capacitance C = dW/dΨ over the range where that relationship can reasonably be approximated as linear.
For our invented one-store system, assume a fixed source potential Ψs, constant supply resistance R and constant outflow E. Inflow is (Ψs − Ψ)/R. Conservation then gives C dΨ/dt = (Ψs − Ψ)/R − E. Every term has units of water volume per time when the units of R and C are chosen consistently.
Setting the time derivative to zero gives Ψeq = Ψs − RE. Subtracting this equilibrium from the changing potential leaves an exponentially relaxing deviation, with time constant τ = RC. This is a derived result for this specified model. It is not a universal formula for any arrangement of roots, stems, leaves and storage tissues.
The model is useful because it makes a prediction about the time course, rather than only the final condition. It can also fail informatively. If measured responses consistently depart from one exponential, investigate changing outflow, more than one storage compartment or state-dependent parameters before treating the discrepancy as random noise.
CHAPTER 8 OF 20 · Build and test a model
8. Read a time constant without turning it into a diagnosis
Choose fictional values R = 0.5 MPa h/mL and C = 4 mL/MPa. Their product is 2 hours. After one time constant, the model retains exp(−1), about 36.8%, of its initial deviation from equilibrium. It has completed about 63.2% of the change. After three time constants, about 5.0% of the deviation remains.
These percentages describe approach to the new equilibrium, not the percentage of all water remaining in a plant. Confusing deviation with total amount would produce a very different and incorrect interpretation. Always identify the reference state when reading a normalised curve.
Now replace the parameters by R = 1 and C = 2 in the same units. Their product remains 2 hours. A normalised relaxation curve alone cannot distinguish these two parameter combinations. To separate them, additional information about transport or storage is required, obtained under conditions compatible with the model.
An estimate of a timescale can still be useful without uniquely identifying its cause. It can help choose a sampling interval or reveal that a supposedly settled measurement was taken too early. The disciplined conclusion is “the response is consistent with this timescale under this model”, followed by the measurements needed to support a more specific physiological interpretation.
CHAPTER 9 OF 20 · Connect physical processes
9. Keep a circuit analogy tied to plant anatomy
An analogy is a translation between relationships, not a claim that the two systems are identical. In a hydraulic circuit representation, a potential difference can be mapped to voltage, flow to current, resistance to resistance and water storage to capacitance. The mapping helps organise equations, but plant tissues still have their own structure, regulation and material properties.
Luo and colleagues’ 2021 primary study developed spatial leaf-hydraulic models with local storage. Its idealised examples included fixed stomatal resistance, and the authors distinguished their modelling predictions from indirect experimental checks. That makes the paper useful for learning how a model’s architecture and assumptions shape its conclusions; it does not establish one parameter set for all leaves.
Before adopting any published circuit, identify the physical meaning of each node and connection. Ask whether storage joins the main pathway directly or through another resistance. Ask whether a boundary is held at a fixed potential or supplies a prescribed flow. Changing either choice can change the governing equation.
The one-store exercise above is deliberately small. It teaches how to derive a model from its connections. Its greatest value may be helping a reader recognise when a more detailed model is necessary, rather than encouraging the same formula to be applied to every new biological diagram.
CHAPTER 10 OF 20 · Connect physical processes
10. Connect water exchange to an energy account
A plant surface exchanges energy as well as matter. For a deliberately simplified steady leaf budget, write Rn = H + L, where Rn is net radiative input, H sensible heat loss and L latent heat loss. Assume storage and other energy transfers are negligible over the chosen interval, with positive terms defined in the directions stated.
Suppose the fictional net input is 250 W/m² and latent loss is 150 W/m². Sensible heat loss must be 100 W/m². Introduce a local linear approximation H = h(Tleaf − Tair), with h = 20 W/m²/K. The required temperature difference is 5 K. If h doubles while the other stated quantities remain fixed, the difference becomes 2.5 K.
This calculation isolates one relationship. A real change in airflow may also change vapour exchange and therefore L. Holding L fixed is an experimental or modelling assumption, not an automatic consequence of moving air faster. The Plant Ecophysiology guide explains how such environmental changes connect to plant performance.
The physical lesson is to ask which terms are controlled and which are allowed to respond. A model prediction based on a fixed latent loss cannot be interpreted as a prediction for freely adjusting stomata and leaf temperature without adding the missing relationships.
CHAPTER 11 OF 20 · Connect physical processes
11. Understand why coupled equations need a shared solution
A coupled system contains quantities that influence one another. A change in water loss can alter cooling. A temperature change can alter the vapour-pressure difference. Conductance affects exchange, while plant water state can influence regulation. Solving each relationship separately with incompatible assumed values does not produce a physically consistent combined result.
In their 2003 coupled-model paper, Tuzet, Perrier and Leuning connected stomatal conductance, photosynthesis, energy balance and water transport. Their methods explicitly stated root geometry and hydraulic assumptions. The lesson for a reader is to inspect those assumptions alongside the equations, rather than treating a successful model as an unrestricted description of every plant.
For a simple invented algebraic analogue, let water loss E = 2 + 0.1θ and temperature excess θ = 10 − 2E, using compatible teaching units. Substitution gives E = 3 − 0.2E, hence E = 2.5 and θ = 5. Both equations are satisfied simultaneously. Taking E = 2 from the first equation while independently assigning θ = 10 would satisfy neither pair consistently.
The example is not a calibrated leaf model. It demonstrates the meaning of coupling and a valuable check: substitute the proposed solution back into every relationship. A plausible-looking temperature is insufficient if the associated water flow violates the other part of the model.
CHAPTER 12 OF 20 · Connect physical processes
12. Distinguish solving an equation from validating a model
Numerical convergence means a calculation has approached a solution under its mathematical rules. Biological validation asks whether the model’s predictions agree with relevant observations. These are separate achievements. A precisely solved model can describe the wrong boundary conditions, omit an important process or use parameters outside their justified range.
For the one-store model, define deviation x from equilibrium. Its equation is dx/dt = −x/τ. A forward-Euler update is xnext = x(1 − Δt/τ). With τ = 2 hours and Δt = 0.5 hours, each step multiplies the deviation by 0.75. After four steps, the numerical fraction is 0.75⁴, about 0.316, compared with the exact two-hour value of about 0.368.
Reducing the timestep can reduce this discretisation error. But if the actual system contains two important storage compartments, improving the solution of the one-store equation will not fix that structural omission. Numerical refinement and model improvement answer different questions.
An especially revealing mistake is choosing Δt = 5 hours here. The multiplier becomes −1.5, producing alternating deviations of growing magnitude although the exact equation relaxes smoothly. Such a pattern is a numerical artefact of this update scheme and step size, not evidence that the fictional plant oscillates. Check the computation before inventing a biological explanation.
A calculated parameter inherits uncertainty from the measurements used to obtain it. If K = Q/ΔΨ, a small uncertain denominator can make K poorly determined. Reporting many digits does not repair that problem. First inspect the quantities from which the estimate was built.
Suppose fictional flow is 2.0 ± 0.2 mL/h and the measured potential difference is 0.50 ± 0.05 MPa. A nominal estimate is K = 4 mL/h/MPa. Using interval extremes gives a range from 1.8/0.55, about 3.27, to 2.2/0.45, about 4.89. This is a conservative arithmetic range under the stated bounds, not a statistical confidence interval.
If the difference were instead 0.05 ± 0.05 MPa, the denominator could approach zero. The same style of finite upper bound would no longer follow. The experiment needs a more informative difference or a better measurement, rather than cosmetic rounding of an unstable ratio.
Also distinguish uncertainty from variation. Measurement uncertainty concerns what is known about a reading. Variation among leaves may be real biological diversity. Repeated readings of one leaf cannot automatically substitute for sampling independent plants when the intended claim concerns a population. State both the experimental unit and the uncertainty interpretation beside the result.
CHAPTER 14 OF 20 · Read the evidence
14. Compare like with like without erasing useful differences
Two studies may report different conductances because they used different definitions, normalisations, states or measurement protocols. Before concluding that the plants differ, align the quantities. Check whether the pathway is whole-plant or leaf-only, whether values are area-normalised, and whether the measurements represent steady conditions or a transition.
This does not mean every difference should be removed. A changed hydration state may be exactly the phenomenon under investigation. The goal is to distinguish intended contrasts from accidental ones. Write down the treatment, the variables held comparable, and the quantities that were allowed to respond.
Consider a fictional experiment comparing two leaves after a sudden light change. One is measured after thirty seconds and the other after twenty minutes. If their responses have different time courses, the comparison mixes identity with observation time. A repair is to record both complete trajectories or compare clearly defined matching phases, rather than simply repeating the mismatched endpoints.
The greenhouse VPD specialist offers a useful related example: equal relative humidity does not establish an equal vapour-pressure difference. Context is part of a physical quantity’s interpretation. A table of conditions is often more valuable than a confident adjective attached to the final graph.
CHAPTER 15 OF 20 · Read the evidence
15. Connect physical state to mechanical response
Transport and mechanics meet when changes in water state alter the behaviour of a living structure. The Plant Biomechanics root examines hydration, material properties, geometry and loading. For biophysical reasoning, the central task is to avoid treating a measured deformation as a direct reading of one hidden property.
Imagine two fictional strips that deflect equally under the same applied force. That equality could arise from matching materials and geometry, or from different combinations that happen to give the same overall response. If one strip is thicker or differently supported, the same displacement does not identify equal material stiffness.
Now suppose a specimen’s displacement increases during an hour. Possible explanations include a time-dependent mechanical response, changing hydration, a shifting support or a changing load. A measurement of displacement alone cannot distinguish all of these. Recording load, time, geometry and water conditions makes the claim more testable.
This chapter connects two types of evidence without reducing one to the other. A hydraulic model does not automatically predict shape, and a bending measurement does not directly measure water potential. To join them, state a relationship linking the physical state to the mechanical property, identify its source or calibration, and test whether that relationship remains appropriate during the experiment.
CHAPTER 16 OF 20 · Read the evidence
16. Design the observation that separates explanations
A useful experiment is chosen for the alternatives it can distinguish. Suppose model A attributes a slow response to high storage and low resistance, while model B attributes it to low storage and high resistance. If both predict the same normalised relaxation curve, collecting that curve more densely may improve the timescale estimate without resolving the mechanism.
Seek a measurement whose predictions differ. In the fictional one-store example, a steady-state flow-and-difference measurement can constrain R under its assumptions. Combined with τ, it permits C = τ/R. An independent storage relationship supplies another route. Neither method should be treated as automatically valid for an arbitrary intact plant: the pathways and conditions must match the quantities in the model.
Plan the comparison before examining the final results. Write each model’s prediction, the observation that would disagree with it, and the resolution needed to separate the predictions. If both alternatives predict differences smaller than the instrument can resolve, the proposed experiment is unlikely to settle the question.
A simple learning activity can use provided fictional data rather than specialised apparatus. Give one group endpoint measurements and another group the same data plus an intermediate potential. Ask which claims each group can defend. The difference in justified conclusions makes the purpose of experimental design visible without requiring a complicated laboratory demonstration.
When an explanation fails, repair the earliest unsupported link. Adding a sophisticated equation after a mislabeled measurement makes the conclusion harder to audit. Use this table to identify whether the problem lies in the quantity, boundary, model or inference.
| Observation or claim | What needs checking | Useful repair |
|---|---|---|
| High flow proves high conductance | Driving difference may also be high | Measure the endpoints and calculate with consistent units |
| Supply equals instantaneous loss | Storage may be changing | Write and check the water balance |
| Slow relaxation proves large storage | Resistance also affects the model timescale | Add an independent constraint on a parameter |
| Equal air humidity means equal demand | Temperature and leaf conditions may differ | Compare the relevant vapour-pressure differences |
| A computed oscillation proves a biological rhythm | Numerical timestep may be inappropriate | Check the solver against an exact or refined solution |
| Equal bending proves equal material | Geometry and supports may differ | Separate structural and material descriptions |
| A precise fit proves the mechanism | Alternative models may fit too | Test predictions under changed conditions |
After selecting a repair, predict what information it will add. “Collect more data” is too vague when every new observation repeats the same limitation. A better instruction is to measure an intermediate potential, resolve the early time course, record leaf temperature, or check the storage term. Each action should address an identified ambiguity.
CHAPTER 18 OF 20 · Diagnose and transfer
18. Transfer check: does the slow leaf store more water?
Two fictional leaves are represented by the same one-store model, with constant R, C, source potential and prescribed outflow during each test. Both have a measured relaxation time of 2 hours. A report says that both therefore have the same water-storage capacitance. A second experiment finds steady resistance of 0.5 MPa h/mL for leaf A and 1.0 MPa h/mL for leaf B. What follows? What changes if the second experiment measured a different pathway?
Answer and reasoning
The first conclusion does not follow from the time constant alone. In the stated model, τ = RC. Using the additional compatible measurements, CA = 2/0.5 = 4 mL/MPa, while CB = 2/1.0 = 2 mL/MPa. The same timescale corresponds to different parameter combinations. These are local capacitances in the model, not total water volumes.
If the resistance experiment measured a different pathway, combining its R with the original τ is unjustified. The calculation can remain numerically neat while losing its physical meaning. The repair is to align the measurement boundary with the model or derive a model that explicitly includes the additional pathway.
Now change another condition: stomatal outflow varies during relaxation. The original constant-E derivation no longer applies unchanged. Measure or model that variation before attributing the response entirely to storage and resistance. Successful transfer means noticing which assumption changed, rather than applying RC because the graph still looks smooth.
CHAPTER 19 OF 20 · Diagnose and transfer
19. Build a learning route from explanation to use
Start with Plant Physiology if the biological pathways are unfamiliar. Follow the transpiration specialist for the water journey and the greenhouse mathematics article for vapour-pressure calculations. Return here when the question becomes which quantities must be connected to test an explanation.
For environmental applications, continue into Plant Ecophysiology. For forces, geometry and deformation, use Plant Biomechanics. These routes serve different questions, and an investigation may need more than one. A physical model is strongest when the biological identity of its components remains clear.
A practical week of study can proceed through four products: a labelled system boundary, a quantity ledger, one worked model and a changed-context critique. Use a spreadsheet to calculate the serial-resistance cases or the exact exponential response. Label all inputs as invented and keep the equations visible. Change one assumption at a time and write what the change means physically.
The aim matches Education: learning that transfers into life: explain a new case with reasons and limits. Explore related disciplines through the Gold Standard master and How X Works hub. The useful outcome is an explanation another person can check, revise and apply appropriately.
CHAPTER 20 OF 20 · Diagnose and transfer
20. Questions and sources for further investigation
Is plant biophysics just plant physiology with equations?
The subjects overlap, but the emphasis differs. Physiology asks how plants function; biophysics gives particular attention to physical relationships, quantities, constraints and measurement. Equations are useful when they clarify those relationships. Adding symbols without improving the testable explanation adds little.
Is the simplest model always the best?
The useful model is the simplest one adequate for the question and evidence. A one-store model may describe a coarse response while missing important spatial differences. More detail is justified when it changes a relevant prediction and can be supported, not merely because it makes the diagram look realistic.
Does a good fit prove the mechanism?
No. Different parameter combinations or different structures can produce similar observations. Check what the data identify, compare alternatives and test predictions under conditions not used merely to tune the fit. A model can be practically predictive while its internal explanation remains uncertain.
Can these calculations determine how much to water a plant?
The invented parameters do not support a watering prescription. Real decisions require plant-specific evidence, substrate conditions, weather, measurement quality and the intended outcome. The calculations teach how to inspect a claim; they do not replace the observations needed to make that decision.
Which primary studies support the research examples?
Sources checked on 24 October 2026 include Tuzet, Perrier and Leuning’s coupled-model study and Luo, Ho, Helliker and Katifori’s spatial capacitive leaf-hydraulics study. Read their assumptions and experimental scope alongside their results. The worked numbers throughout this guide are separate original teaching examples, not reproduced datasets or numerical replications of those studies.