Hydrograph Shape Sampling Validation • rfaR

Purpose

Validate that the hydrograph shape sampling module correctly applies user-specified weights when selecting hydrograph shapes during simulation. This test validates that hydrograph_setup() correctly normalizes weights to probabilities and that sample() using those probabilities reproduces the expected distribution.

Input Data

Seven example hydrographs are loaded via hydrograph_setup() with the following user-specified weights:

hydrograph_names <- c("April 1999", "June 1921", "June 1965 (1hr)",
                      "June 1965 (15min)", "May 1955", "PMF", "SDF")
weights <- c(0.5, 1, 1, 2, 2, 2, 4)
expected_probs <- weights / sum(weights)

Hydrograph Sampling Weights and Normalized Probabilities
Hydrograph	Weight	Expected Prob.	Expected Prob. Fraction
April 1999	0.5	0.04	0.5/12.5
June 1921	1.0	0.08	1/12.5
June 1965 (1hr)	1.0	0.08	1/12.5
June 1965 (15min)	2.0	0.16	2/12.5
May 1955	2.0	0.16	2/12.5
PMF	2.0	0.16	2/12.5
SDF	4.0	0.32	4/12.5

Test

Set up hydrographs with weights, then sample 1,000,000 times and compare the sampled frequencies to the expected probabilities.

set.seed(42)

hydrographs <- hydrograph_setup(jmd_hydro_apr1999,
                                jmd_hydro_jun1921,
                                jmd_hydro_jun1965,
                                jmd_hydro_jun1965_15min,
                                jmd_hydro_may1955,
                                jmd_hydro_pmf,
                                jmd_hydro_sdf,
                                critical_duration = 2,
                                routing_days = 10,
                                weights = weights)

Nsims <- 1000000
hydro_probs <- attr(hydrographs, "probs")
hydroSamps <- sample(1:length(hydrographs), size = Nsims, replace = TRUE,
                     prob = hydro_probs)
sample_freq <- tabulate(hydroSamps, nbins = length(hydro_probs)) / Nsims

Expected vs. Sampled Hydrograph Frequencies (N = 1,000,000)
Hydrograph	Weight	Expected Prob.	Sampled Freq.	Difference
April 1999	0.5	0.04	0.0398	-2e-04
June 1921	1.0	0.08	0.0800	0e+00
June 1965 (1hr)	1.0	0.08	0.0803	3e-04
June 1965 (15min)	2.0	0.16	0.1601	1e-04
May 1955	2.0	0.16	0.1603	3e-04
PMF	2.0	0.16	0.1600	0e+00
SDF	4.0	0.32	0.3196	-4e-04

Acceptance Criterion

Sampled frequencies must be within 1% relative tolerance of the expected probabilities, consistent with expect_equal(sample_freq, hydro_probs, tolerance = 0.01).

Metric	Value
Maximum Relative Difference	0.0042
Tolerance	0.01 (1% relative)
Result	PASS