THE GOLD STANDARD OF HYDROLOGY · SPECIALIST STUDY
Preserve how river errors move together across gauges: calibrate each marginal honestly, model spatial and temporal dependence, test coherent flood fields and refuse unsupported transfer into ungauged regimes
A hydrological correction can be well calibrated at every gauge and still fail the system because independently sampled residuals destroy the dependence between upstream and downstream flows, tributaries, common storms and regional extremes. The gold standard separates marginal error distributions from the dependence model, aligns time and travel lags, uses copulas or resampling only with defensible data support, validates joint exceedances and event coherence, accounts for sparse and changing gauge networks, and limits transfer where catchment similarity, regulation or climate regime is not represented.
This guide belongs to the Hydrology subject guide, the Gold Standard of Education, and the How X Works learning library. Choose a reading route, work the examples, then use the transfer checks to test whether the method survives a changed context.
Choose a reading route
Build the foundations · chapters 1–2
Apply the core method · chapters 3–5
Test evidence and alternatives · chapters 6–7
Handle limits and disagreement · chapters 8–9
Transfer and govern · chapters 10
The whole problem at a glance
| Evidence layer | Question | Failure if ignored |
|---|---|---|
| Marginal | Is each gauge calibrated? | Local bias or spread |
| Spatial | Which gauges err together? | Incoherent flood field |
| Temporal | How does dependence evolve? | Broken event path |
| Network | Which sites are observed? | Unsupported transfer |
Joint decisions need joint predictive evidence.
CHAPTER 1 OF 13 · Build the foundations
1. Define the multisite decision before the model
Core idea. Dependence matters differently for reservoir operation, regional flood warning, ecological flow, insurance and stochastic simulation. The required sites, lead times and joint events follow from the decision.
Method. State the network, forecast horizon, thresholds, routing relations and loss before selecting correlation or copula families.
Worked example. A warning system needs simultaneous threshold probabilities across tributaries, while a water-supply model needs seasonal volume coherence.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. One model may not serve peak timing, low flows and long-term totals equally well.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 1 is complete only when the claimed result survives that review.
Misconception to avoid. Higher spatial correlation is always safer. Risk depends on the decision and network.
Practical evidence check
- Write two decisions using the same gauges and show which joint summaries differ.
- Return to chapter contents
CHAPTER 2 OF 13 · Build the foundations
2. Synchronise observations and legacy simulations
Core idea. Residual dependence is meaningful only after timestamps, units, aggregation intervals, time zones and routing lags are aligned. Misalignment can create or erase correlation.
Method. Version time standards, quality flags and rating curves; test lagged cross-correlation and align by forecast origin rather than hindsight peak matching.
Worked example. An upstream peak leads downstream by six hours; zero-lag residuals appear weakly related until travel time is represented.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. Travel time changes with flow and regulation, so one fixed lag may be insufficient.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 2 is complete only when the claimed result survives that review.
Misconception to avoid. Daily timestamps make sites synchronous automatically. Aggregation and routing still shape dependence.
Practical evidence check
- Shift one series across plausible lags and inspect which dependence claims remain stable.
- Return to chapter contents
CHAPTER 3 OF 13 · Apply the core method
3. Calibrate marginals before dependence
Core idea. Copulas join marginal distributions through dependence; poor local bias, scale or tail models cannot be repaired by a sophisticated copula.
Method. Fit and validate each gauge’s conditional residual distribution, transform to probability-scale values with uncertainty and inspect uniformity by regime.
Worked example. One gauge retains high-flow underdispersion; its pseudo-observations cluster near one and distort the fitted spatial tail.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. Marginal parameters and probability transforms are estimated, adding uncertainty to dependence inference.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 3 is complete only when the claimed result survives that review.
Misconception to avoid. Copulas make marginal calibration optional. They assume the marginal information is meaningful.
Practical evidence check
- Replace one marginal with a misspecified normal model and observe the dependence change.
- Return to chapter contents
CHAPTER 4 OF 13 · Apply the core method
4. Separate spatial from temporal dependence
Core idea. Residuals can correlate across gauges at one time and persist through successive times within each event. Treating all observations as independent inflates evidence.
Method. Model innovations after temporal structure or use a joint spatiotemporal framework, retain event blocks and validate paths as well as snapshots.
Worked example. A storm produces six hours of positive residuals at three gauges; counting 18 independent pairs overstates precision.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. Complex spatiotemporal models can be weakly identified in short records.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 4 is complete only when the claimed result survives that review.
Misconception to avoid. A spatial copula at each hour reproduces whole events. Temporal sequencing remains separate.
Practical evidence check
- Shuffle hours within events and see which diagnostics detect the destroyed path.
- Return to chapter contents
CHAPTER 5 OF 13 · Apply the core method
5. Choose a dependence representation for the dimension
Core idea. Gaussian, t, Archimedean, vine, factor and maximum-entropy copulas encode different symmetry, tail and high-dimensional structures. No family is universally best.
Method. Use hydrological hypotheses and diagnostics to shortlist families, penalise complexity and retain simpler correlation or resampling baselines.
Worked example. A t copula captures symmetric tail co-occurrence, while a vine allows different pair families among upstream and tributary sites.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. High-dimensional vines require many choices and tail parameters are data-hungry.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 5 is complete only when the claimed result survives that review.
Misconception to avoid. The copula with the best likelihood is the true flood mechanism. Fit is conditional and descriptive.
Practical evidence check
- Compare candidate families on held-out events and decision-relevant tail scores.
- Return to chapter contents
Two-gauge example
| Model | Marginals | Joint peaks | Operational meaning |
|---|---|---|---|
| Independent draws | Correct | Too rare together | Understates regional load |
| Perfect correlation | Correct | Too frequent together | Overstates system stress |
| Fitted copula | Correct | Data-supported tail | Candidate model |
| Transferred copula | Unknown | Unverified | Do not certify |
Dependence choice matters most for simultaneous thresholds and routing.
CHAPTER 6 OF 13 · Apply the core method
6. Preserve upstream–downstream and mass constraints
Core idea. Statistical dependence can generate combinations inconsistent with routing, intervening inflow, regulation or conservation even when correlations look realistic.
Method. Check lagged ordering, flow continuity, confluence balance and reservoir rules; couple statistical corrections with routing where required.
Worked example. A simulated downstream peak precedes the upstream release under a purely contemporaneous copula, violating operational chronology.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. Observed data include ungauged lateral inflow and measurement error, so exact algebraic conservation may not hold.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 6 is complete only when the claimed result survives that review.
Misconception to avoid. A correct rank correlation guarantees physical coherence. Dependence summaries do not enforce process constraints.
Practical evidence check
- Route sampled upstream trajectories through a simple network and compare with directly sampled downstream flows.
- Return to chapter contents
CHAPTER 7 OF 13 · Test evidence and alternatives
7. Diagnose tail dependence and joint exceedance
Core idea. System risk often depends on gauges crossing thresholds together. Linear correlation can miss asymptotic or threshold-specific co-occurrence.
Method. Estimate joint exceedance curves, tail-dependence diagnostics and event coincidence with uncertainty; predefine thresholds and compare empirical and simulated counts.
Worked example. Two models share correlation 0.7, but one produces far more simultaneous 1-in-20 threshold exceedances.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. Extreme samples are small, thresholds are uncertain and asymptotic estimates can be unstable.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 7 is complete only when the claimed result survives that review.
Misconception to avoid. Correlation fully determines joint flood risk. That holds only for restricted distribution families.
Practical evidence check
- Construct two copulas with similar rank correlation and compare joint upper-tail probabilities.
- Return to chapter contents
CHAPTER 8 OF 13 · Test evidence and alternatives
8. Validate coherent flood fields
Core idea. A multisite ensemble should reproduce realistic spatial footprints, timing, volumes and recession patterns, not merely pairwise statistics.
Method. Hold out complete regional events, compare affected-site count, peak-lag matrix, regional maxima, volumes and trajectory scores, and inspect ensemble maps.
Worked example. Independent post-processing gives correct coverage at each gauge but almost never produces the observed three-tributary flood footprint.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. Event definitions and missing gauges complicate footprint comparison.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 8 is complete only when the claimed result survives that review.
Misconception to avoid. Passing every gauge independently proves joint calibration. Joint outcomes need joint checks.
Practical evidence check
- Evaluate the same ensemble with marginal coverage and a regional energy or variogram score.
- Return to chapter contents
Coherence certificate
| Claim | Evidence | Boundary |
|---|---|---|
| Marginals valid | Gauge-wise PIT and coverage | Observation error |
| Dependence valid | Pairs, events and tails | Sparse extremes |
| Hydrograph coherent | Lag and volume checks | Regulation changes |
| Transfer supported | Held-out site or basin | Novel climate |
A copula fit is not evidence of physical causation.
CHAPTER 9 OF 13 · Test evidence and alternatives
9. Account for sparse and changing gauge networks
Core idea. Short records, missing periods, relocated gauges and changing rating curves create unequal overlap among site pairs and can bias dependence estimates.
Method. Publish overlap matrices, quality masks and record lengths; use missing-data methods cautiously and test sensitivity to common-period restriction.
Worked example. Two gauges correlate strongly over five flood-rich years but have no overlap during the earlier dry decade.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. Restricting to complete cases can discard information and alter climate composition.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 9 is complete only when the claimed result survives that review.
Misconception to avoid. A large total number of observations means every pair is well supported. Pairwise evidence can be thin.
Practical evidence check
- Re-estimate using full pairwise overlap and a common period, then compare.
- Return to chapter contents
CHAPTER 10 OF 13 · Handle limits and disagreement
10. Represent observation and rating-curve error
Core idea. Shared rainfall products, rating methods and field campaigns can create correlated observation error that is not catchment-process dependence.
Method. Model or bound rating-curve uncertainty, compare alternative observation products and label dependence as residual association unless causes are established.
Worked example. Regional high flows use extrapolated rating curves from one campaign, creating common bias across gauges.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. Separating process and observation dependence may be impossible without auxiliary data.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 10 is complete only when the claimed result survives that review.
Misconception to avoid. All residual correlation is hydrological connectivity. Measurement systems can contribute.
Practical evidence check
- Perturb rating curves jointly and independently and observe the inferred copula.
- Return to chapter contents
CHAPTER 11 OF 13 · Handle limits and disagreement
11. Transfer only with basin and regime evidence
Core idea. Ungauged or short-record sites may borrow dependence from hydrologically similar basins, but distance alone ignores regulation, climate, geology and network topology.
Method. Define similarity features, validate leave-one-site-out, widen uncertainty, flag extrapolation and abstain when the target lies outside support.
Worked example. A nearby regulated tributary has dependence unlike an unregulated neighbour despite similar distance and area.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. Few sites can make similarity models unstable and climate change can invalidate historical analogues.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 11 is complete only when the claimed result survives that review.
Misconception to avoid. Nearest gauge provides the best copula. Geographic proximity is only one feature.
Practical evidence check
- Hold out each gauge in turn and compare transferred joint exceedances with observations.
- Return to chapter contents
CHAPTER 12 OF 13 · Handle limits and disagreement
12. Monitor nonstationarity and network intervention
Core idea. Dependence can change after reservoirs, diversions, urbanisation, gauge technology or climate regimes, even when marginal recalibration appears adequate.
Method. Track rolling rank and tail diagnostics, annotate interventions, validate split periods and version the dependence model and network graph.
Worked example. A new reservoir reduces downstream peak coincidence while upstream marginals remain similar; the old copula overstates system risk.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. Rolling estimates are noisy and apparent change can reflect a few events.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 12 is complete only when the claimed result survives that review.
Misconception to avoid. A copula separates marginals, so it is stationary by construction. Its parameters and family can drift.
Practical evidence check
- Compare pre- and post-intervention joint event diagnostics before pooling.
- Return to chapter contents
CHAPTER 13 OF 13 · Transfer and govern
13. Publish a reproducible joint forecast record
Core idea. A trustworthy multisite product links marginal models, probability transforms, dependence family, parameters, network topology, event definitions, validation and transfer limits.
Method. Version all components, retain random seeds and ensemble members, show marginal and joint diagnostics, protect sensitive infrastructure locations and define retirement triggers.
Worked example. An operator reproduces one regional flood ensemble, traces a tail warning to specific sites and sees why an ungauged branch is excluded.
Changed-context transfer. Changed-context transfer: apply the same reasoning structure to a school or public-service decision. Preserve the formal relationship, then replace the population, authority, privacy duties, evidence and remedy; never copy the domain conclusion.
Limits. Open release may be restricted for critical infrastructure, requiring controlled review.
Evidence and review. For this chapter, keep a dated evidence record: the input, rule or assumption, result, accountable owner and a condition that would change the conclusion. Work an ordinary case and a boundary case, preserve rejected alternatives, and require an authorised independent reader to reproduce where judgement entered. Chapter 13 is complete only when the claimed result survives that review.
Misconception to avoid. A copula parameter table is a complete audit. Joint prediction is a pipeline.
Practical evidence check
- Have an independent team reproduce one held-out event and one abstention.
- Return to chapter contents
Build the specialist decision record
For “The Gold Standard of Hydrology | Multisite Residual Dependence, Spatial Copulas, Coherent Flood Fields, Gauge-Network Sparsity and Transfer”, begin with one bounded decision rather than an abstract claim. Name the affected people, data or institutions; the time window; the authority for acting; and the conclusion the method is allowed to support.
Preserve source material before interpretation. Record definitions, exclusions, transformations and the person responsible for each judgement. A reviewer should be able to separate an observed fact from an inferred result, a policy choice and a presentation shortcut.
Work one ordinary case in full. Show the input, every material rule or calculation, the output and the action that follows. Then write what the same evidence would mean under the simplest credible alternative method.
Keep an explicit unresolved state. Missing, protected, conflicting or unevaluated information must not become zero, false, consent or absence merely because a form expects one value.
Finish the record with a claim ledger. For every sentence that could drive action, identify its supporting source, inferential step, confidence or legal status, intended reader and expiry condition. This keeps descriptive evidence, formal consequence, professional judgement and institutional authority from collapsing into one impressive-looking conclusion.
Run the boundary-case audit
Select a case near the method’s domain edge: a missing input, a disputed definition, an endpoint, a sparse group, a resource limit or a legal deadline. Boundary cases reveal the contract that routine examples can conceal.
Change one assumption at a time and retain every output. Mark what stays stable, what reverses and what becomes undefined. If several assumptions move together, report that dependence instead of attributing the change to one convenient cause.
Search deliberately for a counterexample to the strongest claim. A method that works on the published illustration can still fail under a permitted input, an alternative grouping, a changed unit or an inaccessible evidence pathway.
Translate the counterexample into plain language: who would receive a different decision, what evidence would be hidden, and which safeguard or narrower claim repairs the problem.
Include an adversarial but plausible case. Let a missing group, ambiguous identifier, extreme weight, conflicting rule or urgent deadline interact with normal system incentives. The aim is not theatrical failure; it is to learn whether ordinary pressure makes the method conceal uncertainty, shift burden or reward a misleading shortcut.
Test independent reproduction
Give a reviewer the bounded inputs, public method contract and permitted access, but not the desired answer. Ask them to reproduce the result and identify every place where judgement, local knowledge or unavailable evidence was required.
Compare the reviewer’s intermediate states, not only the final number or label. Matching outputs can hide different denominators, inference paths, missing-data treatments or authorities that will diverge on the next case.
Where reproduction fails, classify the cause: ambiguous definition, undocumented transformation, incompatible implementation, inaccessible source, calculation error or legitimate alternative interpretation. Each class needs a different remedy.
Retain the disagreement and resolution. A corrected record should show what changed, why it changed and which prior outputs or decisions must be revisited.
Run the reproduction on a changed but structurally similar case after the first discrepancy is repaired. A method is not transferable merely because its authors can explain the original example. The new case tests whether the public contract carries the reasoning or whether hidden expertise is still doing essential work.
Design correction, challenge and remedy
Assign a visible route for correcting inputs, challenging assumptions and reviewing the application of the method. Technical accuracy without a usable remedy leaves the affected person carrying an error they cannot contest.
Separate urgent live-case repair from later system learning. A current decision may need pause, reassignment, fresh evidence or a human review before the organisation finishes its longer investigation.
Protect confidential and privileged material during review, but do not use confidentiality as a reason for unrecorded authority. Layer public explanation, protected oversight and individual reasons according to lawful access.
Verify that correction reaches cached results, derived records, dashboards and future decisions. Closing a complaint is not enough if the old conclusion continues to travel.
Measure remedy from the affected person’s position. A revised database row is useful, but it is incomplete when a hearing, service allocation, published claim or downstream model still relies on the old result. Record notification, restored opportunity, compensation or procedural repair where the governing framework provides it.
Set review and retirement triggers
Review after a change in definitions, population, law, source access, measurement scale, software, vocabulary or incentives. A stable formula can become invalid when the world or the meaning of its inputs changes.
Monitor leading signals as well as final outcomes: coverage gaps, timeouts, overrides, complaints, disagreement, missingness and boundary frequency. These often reveal failure before a headline metric moves.
Predefine thresholds for recalibration, suspension and retirement. Do not force a method to answer after its assumptions fail merely because replacing it is inconvenient.
At retirement, preserve the versioned contract, evidence, known limitations and migration decision. State which historic results remain interpretable and which require warning, restatement or withdrawal.
Schedule one prospective review before confidence fades or controversy forces it. Assign named owners for evidence, implementation and rights impact, and invite at least one reviewer outside the design team. The review should be able to narrow the claim, change the method or stop use without requiring proof of catastrophic failure.
Questions readers ask
Why is gauge-by-gauge calibration insufficient?
- It does not determine whether errors and extremes occur together across the network.
What does a copula do?
- It represents dependence separately from the marginal distributions.
Does high correlation prove coherent floods?
- No. Timing, tail co-occurrence, routing and physical constraints also matter.
Why hold out complete events?
- Random time splits leak dependence and do not test regional flood footprints.
Can dependence be transferred to an ungauged site?
- Only with similarity evidence, held-out validation and explicit uncertainty or abstention.
What changes can invalidate a model?
- Reservoirs, diversions, land use, climate, gauges, rating curves and upstream model versions.
Practise, verify and continue
Choose one consequential case. Reproduce the complete method, work one ordinary and one boundary case, test the strongest rival explanation, and state the evidence that would reverse or suspend the result. Build an evidence pack containing original inputs, definitions, calculations or rule steps, alternatives rejected, uncertainty, reviewer comments and the final action. Give the pack to a reader who did not design the method. Ask that reader to reproduce the result, identify one missing assumption and construct a case that would make the method return unresolved. Revise the public explanation only after the evidence pack and the operational process agree.
Move the method into an education or public-service setting. Keep the reasoning structure while replacing domain authority, population, privacy, evidence and correction routes. Start with a low-stakes historical example, then compare the specialist method with a transparent baseline. Check whether the same people, records and decisions remain inside scope; whether the scale or vocabulary still means the same thing; and whether a person affected by error can understand and challenge it. Pilot under human review, record reversals and burden, and stop the transfer if the formal resemblance hides a different ethical or legal problem.
Related eduKate reading
State-dependent residual models and transfer limits
Posterior predictive checks and event dependence
Bayesian total-error analysis and rating-curve uncertainty
Authoritative sources and scope
Chen et al., Copula-Based Multisite Streamflow Simulation
Hao and Singh, Multisite Streamflow Dependence With Maximum Entropy Copula
A Framework of Dependence Modeling and Evaluation for Multivariate Hydrology
Synthetic Simulation of Spatially Correlated Streamflows
Wang and Shen, Vine Structure for Temporal and Spatial Streamflow Dependence
This article is multisite hydrological uncertainty education, not an operational flood warning or infrastructure certification. Examples are original teaching cases. Joint models remain conditional on observations, rating curves, marginal calibration, network structure and historical regimes. Use qualified local expertise, protect sensitive infrastructure data, validate complete events and abstain from unsupported transfer.