Updated Hypothesis Document

Buffalo River Watershed — Running Hypothesis Document

Last updated: 2026-06-10 (Day 102 — recession-only day appended; document carried forward complete from Day 102 rebuild) Events observed to date: 19 confirmed rainfall-runoff events + multiple sub-threshold non-detections + 3 clean recession-only baseline days (Days 95, 97, and Day 102 high-flow recession tail of Event 19)

Study Context

• 90-day MRMS/USGS empirical study, Buffalo River HUC8 11010005 (Feb 27 – May 29 nominal; extended through Day 102 = Jun 10 as bonus calibration data).
• 7 gauges (5 mainstem + 2 tributary signal-separators); 37 HUC12s including 8 ungauged below Harriet (39% of basin area is St. Joe contributing).
• Pre-study calibration anchors: Ponca + Pruitt rating curves recalibrated Feb 17, 2026.
• Local-knowledge calibration targets in active use: Hailstone pool-and-drop pre-fill; Pruitt low-flow river-right channel scour; Ponca historical ~48-hr rating-curve lag during rises; Hailstone recreational thresholds (3.7/4.4/5.9 ft); upper-Richland recreational thresholds (3.2/4.0/6.0 ft); Boxley-Ponca section thresholds (3.2/3.7/5.0 ft).
• Days 91–102 are post-study bonus data — calibration-grade, with multiple major rainfall-runoff events captured.

Events

• Event 1 (Day 5–6, Mar 5–6): First confirmed event. Headwater-concentrated storm, Boxley HUC12 1.038" with 0.751"/hr peak (study-record intensity to that date). Dry antecedent (~0.4" 7-day). Boxley +0.87 ft biphasic (slow fill 8 hr → rapid 0.46 ft/hr pulse), peak at 15:15 with ~11 hr lag from peak rain — classic "pool and drop" signature. Ponca 84 → 179 cfs (+113%), 4-hr Boxley→Ponca lag. Pruitt +0.63 ft / +144 cfs (+231%) arriving Day 6, 15-hr Ponca→Pruitt lag at ~150 cfs flow (~0.46 m/s). Pruitt rating curve shifted mid-event (a real USGS update). St. Joe non-response — first evidence that 0.5" distributed across 1342 km² is sub-threshold. Transfer ratios: Boxley 0.84 ft/in (dry); Pruitt brief Optimal crossing.

• Event 2 (Day 7–8, Mar 7–8): Watershed-wide major storm, 1–2.5" basin-wide. Ponca zone hotspots Whiteley 2.55"/0.96"/hr, Pruitt zone Cove 2.43"/0.74"/hr. Wet antecedent from Event 1 (~1.0–1.5" 7-day Ponca/Pruitt). First flood-threshold crossings of study: Ponca 2120 cfs (>1600), Pruitt 2610 cfs (>2000, 3.5 hr above). St. Joe 8.01 ft / 3230 cfs at 03:30 Day 8 (4.60 ft rise, 23.5-hr lag from QPE centroid). Harriet 6.83 ft / 3050 cfs (0.99 ft/hr peak rise, fastest of study). Bear Cr first sustained response: 1.2" → +0.59 ft, 11-hr lag. Transfer ratios: Boxley 0.71 ft/in (wet); Ponca ~1030 cfs/in (wet); Richland 1.39" → +1.39 ft (1.0 ft/in wet). Propagation cascade: Boxley→Ponca 4–5 hr, Ponca→Pruitt 5.75 hr at 1500–2000 cfs (~1.2 m/s), Pruitt→St. Joe 14.25 hr, St. Joe→Harriet 7 hr. Total Ponca→Harriet 27 hr at flood flow.

• Event 3 (Day 11, Mar 11): Stacked on Event 2 recession. 0.5–0.8" upper-basin focus on already-elevated baseflow. Boxley +0.66 ft from 0.811" (0.81 ft/in wet, ~10.5 hr lag — Boxley lag ≈ constant across antecedents in 1.0"-class events). Ponca 197→547 cfs (+178%). Pruitt biphasic: local +0.10 ft then propagated +1.32 ft, peaking 5.93 ft / 857 cfs. St. Joe 6.02 ft / 1420 cfs Day 12 (+1.00 ft from trough), 25-hr Ponca→St. Joe lag. Harriet 5.42 ft / 1510 cfs Day 12, 33-hr Ponca→Harriet. Ponca→Pruitt 7.5 hr at ~500 cfs (~0.93 m/s) — confirms velocity ∝ Q^0.35 trend.

• Event 4 (Day 35–37, Apr 3–5): Three-pulse compound: lower-watershed primer Apr 3 (0.4–0.7"), then watershed-wide 1.2–1.95" Apr 4 (Shop Cr 1.95" hotspot). Wet/wet antecedent. Boxley +2.05 ft from 1.57" = 1.30 ft/in (wet, highest observed) — confirms pool-fill primer mechanism. Pruitt 8.55 ft / 2610 cfs (transient flood crossing, 3.22 ft rise on Pruitt day). St. Joe 7.40 ft / 2630 cfs at 12:30 Day 37 (4.0 ft rise, ~33.5-hr lag from QPE centroid — distributed-forcing slower than concentrated). Harriet 6.83 ft / 2720 cfs at 19:15 Day 37 (3.30 ft rise). Bear Cr 4.6 → 203 cfs (+4300%), 0.97 ft/in transfer. Pruitt→St. Joe 12 hr; St. Joe→Harriet 7 hr. Sustained Optimal at Pruitt ~135 hr; St. Joe/Harriet Optimal duration ~21 days (with stacking).

• Event 5 (Day 41–42, Apr 10–11): Localized two-zone event. Harriet zone 0.6–1.1" (Water Cr 1.10") → Harriet +0.10 ft / +73 cfs at +1–5 hr lag (first clean Harriet-zone-only pair). Boxley zone 0.55" → +0.20 ft delayed (0.36 ft/in, moderate antecedent) → propagated through Ponca +20 cfs and Pruitt +0.12 ft / +34 cfs, with Boxley→Pruitt 15-hr at ~185 cfs. Cove Cr 0.66" produced compound Pruitt response. Transfer-ratio non-linearity confirmed: 0.36 ft/in at 0.55" zone QPE vs 0.71–0.84 at 1.0"+ — Boxley has threshold-type response.

• Event 6 (Day 46–47, Apr 15–16): Two-pulse Apr 15-16 (Richland-Cave bullseye) + Apr 18 (upper basin 0.4–0.7"). Cleanest propagation cascade of early study. Pulse 1: Boxley peak ~12:45 → Ponca 17:00 (+4.25 hr) → Pruitt 07:15 Day 47 (+14 hr Ponca→Pruitt at low flow, velocity-discharge curve confirmed) → St. Joe peak 08:00 Day 48 (~43 hr from forcing) → Harriet 18:00 Day 48 (~53 hr). Pulse 2 propagated separately with similar cascade. Boxley 0.85"-cumulative → +0.48 ft, 0.56 ft/in moderate antecedent (intermediate between dry 0.36 and wet 0.84). Pruitt briefly re-entered Optimal.

• Event 7 (Day 54–56, Apr 24–26): Tier 2 event after 6-day dry spell. Pruitt zone 1.353" (Cove 1.34", Hoskin 1.36", Flatrock 1.40"). Dry-to-moderate antecedent (0.30–0.70" 7-day). Cascade: Boxley peak 3.21 ft Day 54 16:30 → Ponca 347 cfs 18:15 (+1.75 hr) → Pruitt 5.31 ft / 612 cfs 03:00 Day 55 (~10.5 hr from Ponca) → St. Joe 5.99 ft / 1460 cfs 17:15 Day 55 (+14.25 hr) → Harriet 5.52 ft / 1610 cfs 02:00 Day 56 (+8.75 hr). Total Boxley→Harriet 33.5 hr. Transfer ratios: Boxley 0.82 ft/in (dry, but high-intensity bursts overrun pool storage); Bear Creek first true flash response: 0.99" → +0.27 ft / +18 cfs, 26-hr lag (long dry-antecedent lag at Bear Cr). Richland 0.98" → +1.18 ft (did not reach 3.2 ft).

• Event 8 (Day 59–62, Apr 28–May 1): Two-pulse Tier 2-3 major event. Pulse 1 (Apr 27 evening): lower-watershed primer 0.5–0.85" on dry antecedent. Pulse 2 (Apr 28): watershed-wide 1.0–1.7" peak hour 14:00–16:00 CDT. Saturated antecedent post-Pulse 1. Spatial extremes: Falling Water 1.40"/0.99"/hr; Hoskin 0.81"/hr. Richland +2.90 ft in 60 min at 18:30 (largest hourly rise of study — "wet/saturated + high intensity" mode), peak 4.89 ft (near Upper Richland Optimal 4.0). Boxley 5.50 ft Hailstone-Optimal+ for ~32 hr (exceeded local "rarely >24 hr" rule). Pruitt 5.98 ft / 893 cfs at 10:15 Day 60 (1.42 ft/in wet). St. Joe 6.96 ft / 2260 cfs at 02:00 Day 61 (lag ~30–35 hr, 2.1 ft/in wet cumulative). Harriet 5.52 ft / 1610 cfs at 02:00 Day 61 (1.4 ft/in wet, 9-hr St. Joe→Harriet at high flow). Bear Cr +1.49 ft / +525 cfs in 1 hr (+0.27→3.88 ft, 26-hr lag — Bear Cr long-lag character on saturated soil). Total Boxley→Harriet 52 hr secondary cascade. Recession sequence at St. Joe: 33→16→10→6.6→5.8 cfs/hr per 24-hr window (geometric decay ratio ~0.6–0.9).

• Event 9 (Day 66–68, May 5–7): First clean dry-antecedent calibration event after 6-day dry spell. Single-pulse Day 66 morning, Boxley 1.23"/0.53"/hr. Dry antecedent (7-day 0.07–0.30"). Boxley +0.87 ft, 2.25-hr lag (high-intensity bursts overrun pool storage in dry conditions), transfer 0.71 ft/in (dry, intensity-elevated). Bear Cr 0.92" → +0.32 ft / +34 cfs, 9-hr lag dry (vs 26-hr wet) and 0.35 ft/in (vs 1.45 ft/in wet) — confirms ~4× damping under dry. Richland 1.04" → +0.26 ft, transfer 0.25 ft/in dry (vs 1.0+ wet). Pruitt 0.45" → +0.65 ft / +194 cfs, transfer 1.44 ft/in dry — IDENTICAL to wet (~1.4). Pruitt antecedent-independence finding: confirmed. Cascade: Boxley→Ponca 3 hr (dry slower), Ponca→Pruitt 10.75 hr, Pruitt→St. Joe 20.5 hr dry (vs 14 hr wet, ~46% slower). St. Joe→Harriet 10.25 hr — antecedent-insensitive. Total Boxley→Harriet 44.5 hr dry (~35% longer than 33 hr wet).

• Event 10 (Day 71–74, May 10–13): Tier 2 multi-zone with concentrated Richland forcing. Falling Water 1.40"/1.08"/hr, Boxley 1.32"/0.54". Moist antecedent (~1.0–1.5" 7-day). Richland +1.91 ft / 0.74 ft from low (1.34 ft/in moist + high-intensity — HIGHER than wet 1.0), 4.5-hr lag. Briefly reached Low-Floatable Upper Richland. Boxley 0.67 ft/in (moist, with sustained-rain pool-fill pattern: gradual 4 hr + step +0.63 ft at 21:45). Bear Cr 0.054 ft/in (sub-threshold on still-dry soil, the +0.06 ft / +1.4 cfs micro-response). Double-peak structure at St. Joe: Primary 4.66 ft / 646 cfs at 12:30 Day 72 (Richland-tributary-driven, peaked 1.5 hr BEFORE Pruitt — first explicit demonstration of tributary dominance over mainstem); Secondary 4.53 ft / 584 cfs at 13:15 Day 73 (Pruitt-mainstem-driven, 23.25-hr Pruitt→St. Joe at low-mid flow). Harriet primary 4.41 ft 03:00 Day 73 (14.5 hr St. Joe→Harriet); secondary 4.41 ft 01:45 Day 74 (12.6 hr, 76% cfs retention vs 85% primary).

• Event 11 (Day 80–82, May 19–21): Major watershed-wide flood-scale forcing on extreme-dry antecedent. DRIEST antecedent of any major event (7-day 0.07–0.30"). Pruitt zone 1.93" (highest of any zone — first Pruitt-dominant event), Ponca 1.52", Boxley 1.16", St. Joe 1.60", Bear Cr 1.11" (spatial minimum). Direct-forcing dominance — every zone simultaneously: Boxley first rise 10:15 CDT, Ponca 10:00, Pruitt 09:00, St. Joe 10:00, Harriet 09:15 (parallel direct responses, not sequential cascade). Boxley +0.72 ft (0.62 ft/in dry); Pruitt +1.17 ft / +297 cfs (0.61 ft/in dry — confirms Pruitt dry transfer ratio); St. Joe 5.22 ft / 949 cfs (1.09 ft/in dry — surprisingly close to wet, 37-hr lag from peak QPE); Harriet 5.00 ft / 1130 cfs (0.76 ft/in dry, 52-hr lag — longest peak-to-peak lag of study); Richland +1.02 ft (0.73 ft/in dry); Bear Cr near-total non-response (+0.17 ft / +10.8 cfs from 1.11"!) = 0.09–0.17 ft/in dry — Bear Cr structurally damped under dry. St. Joe→Harriet 11.25 hr (dry-antecedent lower-mainstem signature). Total cascade Boxley→Harriet ~36 hr.

• Event 12 (Day 83–85, May 22–24): Moist-antecedent spatially-shifting multi-day convective. Day 83 lower-watershed-heavy (Bear Cr 0.685", Richland 0.621", Harriet 0.563"); Day 84 shifted NW to Pruitt/Whiteley/Cave (Pruitt 0.319"). Moist antecedent (1.0–2.5" 7-day). Richland 0.443" + 0.036" → peaked 1.85 ft. Bear Cr two-peak hydrograph: D83 +0.05 ft (~30 min lag) + D84 +0.15 ft delayed (37-hr lag from D83 QPE) = 0.25 ft/in total (moist anchor). Pruitt two-mode response confirmed: 6.5-hr surface peak D84 + 19.75-hr karst/subsurface secondary peak D85 — consistent with local-knowledge karst losses between Ponca and Pruitt. MAJOR FINDING: moist Event 12 transfer ratios LOWER than dry Event 11 at all gauges (Boxley 0.0 vs 0.62; Pruitt 0.17 vs 0.61; Richland 0.35 vs 0.73). Forcing magnitude and spatial coherence matter more than antecedent moisture below threshold — distributed sub-1.0" forcing produces diffuse non-coherent hydrographs even on wet soils.

• Event 13 (Day 87, May 26): Richland-bullseye event — Falling Water 0.942"/0.466"/hr, Headwaters Richland 0.695". Moist antecedent (1.77–1.97" Richland 7-day). Richland +0.74 ft from low / 0.90 ft/in (moist, event-scale anchor), 3-hr lag — first event-scale Richland calibration of study. Bear Cr +0.36 ft / +31.2 cfs, 9.5-hr lag, 0.94 ft/in (moist-saturated) — flagged anomaly at the time, now contextualized as wet-mode behavior. Boxley non-detection (0.282" sub-threshold despite moist antecedent — pool-fill mechanism intact). Single-day event; recession Day 88.

• Event 14 (Day 88, May 27): Upstream-shifted bimodal forcing. Bear/Lick Cr afternoon pulse + heavy Boxley/Ponca/Little Buffalo overnight (Boxley 0.706"). Carryover from Event 13: Richland peak propagated to St. Joe morning Day 88 at 4.25 ft / 461 cfs (Richland → St. Joe 19.5 hr; St. Joe → Harriet 12 hr to 4.29 ft / 581 cfs). St. Joe Richland-source transfer ~0.37 ft/in. Event 14 Boxley response began late evening (+0.04 ft, peak pending). Transitioned directly into Event 15.

• Event 15 (Day 88–89, May 28–29): MAJOR WATERSHED-WIDE FLOOD-SCALE EVENT — FIRST flood-threshold-exceeding mainstem peak of study. Largest 24-hr basin QPE of study. Ungauged hotspot Long Cr 6.087" with 1.584"/hr (approaching March 2024 reference event scale). Pruitt zone 2.60", St. Joe 2.14", Bear 2.63", Richland 2.63", Harriet 2.67". Moderate-to-wet antecedent. Pruitt 8.55 ft / 2610 cfs (>flood 2000) bimodal (Mode 1: 13:45 from local Cove/Hoskin/Flatrock; Mode 2: 18:00 from Ponca-zone backwater) — 1.9 ft/in moist. Bear Cr 6.30 ft / 3090 cfs bimodal, 1.53 ft/in (wet/stage-dependent). Richland 8.89 ft (near Upper Richland Optimal 4.0 by gauge convention), 2.69 ft/in — channel-geometry nonlinearity once out-of-bank. St. Joe 15.05 ft / 14,500 cfs at 01:30 Day 90 (~1.8× flood threshold), 11-ft rise, ~3 ft/in event-scale. Harriet 14.24 ft / 15,000 cfs at 07:30 Day 90 (~1.6× flood threshold), +5.69 ft in 7.5 hr (1.68 ft/hr peak). Pruitt→St. Joe ~11.75 hr, St. Joe→Harriet 6 hr at flood flow. Approaches March 2024 reference event magnitude but distributed forcing (not Richland-only) produced same St. Joe stage from less concentrated rain.

• Event 16 (Day 90, May 30): Late-afternoon Cave Cr / Big Cr / Bear Cr / Spring Cr convective sub-event on saturated antecedent (7-day 3.5–5.1"). Cave 0.88"/0.85"/hr, Bear 0.59"/0.45"/hr. Bear Cr +1.23 ft / +956 cfs in 2 hr, 4-hr lag, 2.08 ft/in — highest Bear Cr transfer in study. Refined finding: Bear Cr transfer is stage-dependent, not purely antecedent-dependent — pre-event low stage (3.54 ft, steep rating-curve zone) amplifies ft-per-inch response relative to elevated-stage conditions. Signal separation textbook: Richland gauge flat → St. Joe rebound confirmed as Cave/Big Cr / mainstem-misc source. St. Joe rebound peak D92 01:15 at 9.84 ft / 5550 cfs (~9.25-hr lag, 2.5 ft/in wet-antecedent for concentrated sub-zone forcing). Harriet rebound D92 07:00 at 9.02 ft / 6000 cfs (5.75-hr St. Joe→Harriet propagation — second confirmation of ~6-hr high-flow timing).

• Event 17 (Day 93–94, Jun 1–2): Wet-antecedent multi-cell convective. Cove 1.02"/0.72"/hr, Falling Water 0.93"/0.83"/hr, Headwaters Bear 0.55"/0.41"/hr, Dry Cr 0.80". Very wet antecedent (4–5" 7-day Cove, Falling Water; Bear ~4.15"). Bear Cr was the cleanest pair: 0.51 ft/in, 4-hr lag from 3.07 → 3.38 ft — confirms stage-dependence finding (mid-stage on recession from prior pulse → lower ratio than Event 16's near-baseline 2.08). Pruitt non-response to its zone's 1.02" Cove Cr bullseye — most striking finding, indicating Pruitt's response dominated by upstream routing rather than local-zone forcing when mainstem is in active recession. Richland near-zero rise despite 0.93" Falling Water — saturated channels above gauge had absorbed pulse. Major refinement: Bear Cr wet-antecedent transfer ratio range 0.5–2.0 ft/in depending on initial stage (rating-curve steepness near baseflow vs. floodplain flattening at high stage).

• Event 18 (Day 96–97, Jun 4–5): Late-evening upper-basin convective pulse — Boxley zone 0.302"/0.246"/hr on drying antecedent (1.85" 7-day, down from 2.83"). Sub-threshold confirmed: zero Boxley response through end of Day 97. Confirms ~0.3" insufficient on drying antecedent — pool-fill threshold not crossed. Cleanest in-study example of USGS rating-curve revision artifact: Bear Cr 60.0 → 71.7 cfs step at 18:30 with zero height change and zero zone QPE — discrete cfs-table update applied mid-day. Validates local-knowledge guidance: height > cfs for cross-event comparison; post-major-event field visits produce visible step changes.

• Event 19 (Day 98–101, Jun 6–9): MAJOR UPPER-BASIN MULTI-PULSE EVENT — second flood-scale event of study. Pulse 1 Day 99 AM (Boxley 2.97 ft from 0.532"/hr Terrapin Branch peak, dry-leaning antecedent 0.45–1.2"). Pulse 2 Day 99 PM (smaller, sustained 4-ft+ baseline). Pulse 3 Day 100: STUDY-RECORD Boxley 5.50 ft (Hailstone Optimal+, exceeded all prior peaks), from 0.699"/hr peak in HUC 0201, 3.7-hr peak-to-peak lag, 2.28 ft/in (very wet, intensity-elevated — highest Boxley transfer of study). Ponca 1920 cfs (>1600 flood threshold) at 15:30 Day 100, ~45 min above flood. Pruitt 7.98 ft / 2130 cfs (>2000 flood threshold) at 21:45 Day 100, ~4–6 hr above flood — second flood-threshold event at Pruitt of study. Day 101 propagation peaks: St. Joe 7.63 ft / 2900 cfs at 10:30 (36% of flood threshold) — 25–26 hr Day 100 peak QPE → St. Joe peak; Harriet 6.67 ft / 2860 cfs at 18:45 (30% of flood threshold) — 8.25 hr St. Joe→Harriet. CRITICAL CROSS-EVENT FINDING vs Event 15: upstream-only forcing → upstream-only flood. Watershed routing/storage attenuated flood pulse to ~30–36% of flood threshold at mid/lower mainstem. Pruitt flood was ONLY because of constructive interference between upstream cascade and local Pruitt-zone rain. Bear Cr only +0.06 ft (signal separation: Harriet rise = pure mainstem propagation). Recession tail (Day 102, see Recession Behavior): St. Joe ~32 cfs/hr, Harriet ~37 cfs/hr falling, both still mid-Optimal — Event 19 delivered a multi-day Optimal window across the lower mainstem despite upstream-only flood.

Recession Behavior

• Two-phase recession universally observed: fast interflow phase (first 24–48 hr post-peak) → slow groundwater/baseflow phase. Geometric decay with daily ratio ~0.5–0.9 per 24-hr window.
• St. Joe canonical sequence (Event 8 wet): 33 → 16 → 10 → 6.6 → 5.8 cfs/hr over Days 1–5 post-peak. (Event 9 dry: 6.96 → 5.1 → 3.2 → 1.1 → 0.78 cfs/hr — faster fractional decay under dry, smaller absolute slope.) Event 19 tail (Day 102, ~Day 1–2 post-peak): St. Joe falling ~32 cfs/hr (2000→1270 over 22.75 hr) — sits between Event 8 wet Day-1 (33) and Day-2 (16), appropriate placement on fast interflow limb.
• Harriet sequence: tracks St. Joe at high flow; St. Joe and Harriet converge to near-identical absolute recession rates once propagation pipeline empties (Days 95, 97 baselines confirmed; Day 102 high-flow analog: St. Joe ~32 cfs/hr vs Harriet ~37 cfs/hr — near-identical even at elevated ~1300–2500 cfs, confirming convergence holds across the flow range once the pulse has passed).
• Boxley pool drain: ~4 days for full drain from major-event peak to pre-event baseline. Slow seepage phase ~0.005–0.015 ft/hr. Pool-fill timescale (dry → response): ~36 hr for first 0.5" pulse to register at gauge. Day 102 confirmation: Boxley 3.22→2.87 ft (~0.015 ft/hr) on Day 4 of Event 19 drain from the Day 100 5.50 ft study-record peak — consistent with ~4-day drain timescale.
• Bear Cr: flashy recession, 90%+ depletion in 4 days from major peak. Recession constant k ≈ 0.017–0.024 /hr at typical wet-to-mid-flow. Day 102: Bear Cr drained 37.1→28.5 cfs smoothly with no discrete table steps — post-Event-15/16/17 rating revisions (Days 96, 100/101) appear stabilized.
• Richland Cr: very flashy, returns below pre-event baseline within 3–4 days post-major-peak (no sustained interflow contribution). Day 102 confirmation: Richland 1.45→1.32 ft, back below its pre-Event-19 baseline within ~2 days of the event's negligible Richland response.
• Day 65 "dry-spring baseflow" anchor was a moment in continuing decline, not a true floor. By Days 75–79, all gauges had dropped below Day 65 values during extended dry stretches. No stable seasonal baseflow floor observed in 90-day window; recession is continuous and asymptotic, not floored.
• Day 95 and Day 97 baselines: cleanest recession-only days of study, useful reference for "pipeline-empty" state. St. Joe and Harriet falling at near-identical 0.020–0.022 ft/hr / ~10–15 cfs/hr = pure zone-baseflow drainage. Day 102 adds a HIGH-FLOW recession-only analog (pipeline still partly full, ~1300–2500 cfs): falling rates ~32–37 cfs/hr, ~10× the pipeline-empty baseline rate — quantifies the fast-interflow vs baseflow rate contrast.
• Sub-threshold pulses produce delayed subsurface baseflow support: Day 78 finding (Bear Cr held nearly flat after sub-threshold 0.11" on saturated soils while other gauges dropped). Confirmed at multiple subsequent gauges.
• Pre-event baseline overshoot: after major events, Pruitt and Richland routinely recess BELOW pre-event values during the slow phase, indicating residual storage from pre-event period drained out alongside event flow.

Rating Curve Observations

• Ponca + Pruitt calibrated Feb 17, 2026, two weeks pre-study — both initial baselines reliable.
• Pruitt mid-event rating shift Day 5/6 (Event 1): CFS dropped 67.9 → 50.9 with stable height — first observed in-study USGS update.
• Pruitt low-stage CFS unreliable (height < ~3.7 ft, CFS < ~150). Local-knowledge channel scour shifts flow to river right; gauge on river left under-reads. Persistent Ponca > Pruitt CFS at low flow throughout dry stretches — NOT a karst loss signal at low flow. Height is the reliable Pruitt metric below ~150 cfs.
• Ponca rating-curve revision Day 76: 154 → 91 cfs step at 11:00 with smooth Boxley recession upstream and zero QPE = field calibration update. Pre-Day-76 Ponca low-flow CFS over-reported by ~30–40%. All pre-Day-76 Ponca CFS values flagged accordingly.
• Bear Cr rating-curve revision Day 96: 3.2 → 21 cfs step with zero height change — confirms post-Event-15/16/17 field visits applied. Bear Cr post-Day-96 CFS values are calibrated baseline.
• Bear Cr second revision Day 100/101 indicated (60.0 → 71.7 cfs Day 96 evening). Treat as ongoing calibration improvement. Day 102 update: Bear Cr cfs now steps down smoothly (37.1→28.5) with no discrete jumps — revisions appear stabilized.
• High-stage rating curves performing well: Event 15 Pruitt 8.55 ft → smooth height-discharge tracking. Event 19 Day 100 flood peaks tracked cleanly. Day 102 recession limbs at all gauges tracked smoothly height-vs-discharge with no artifacts.
• Height comparison across gauges is meaningless — each gauge's rating curve is location-specific. Event 19 confirmation: Pruitt 7.98 ft = 2130 cfs (Day 100); St. Joe 7.63 ft = 2900 cfs (Day 101); same approximate height, very different discharges.
• CFS accuracy improves during active weather (field crew visits triggered by events). The study's later cfs values are higher-confidence than early-study values.

Cross-Event Findings

Detection thresholds (Q1) — zone-averaged QPE required for measurable mainstem rise: - Boxley: ~0.5" wet/moist; ~1.0" dry (pool-fill barrier); sub-detection at 0.3" even wet, 0.5" even dry-moderate. - Ponca/Pruitt: ~0.3" wet; ~0.5" dry; Pruitt response dominated by upstream routing — local-zone forcing produces small response even at 0.8–1.0" if no upstream signal. - St. Joe: ~0.5" zone-avg with concentrated cells; ~0.3" if spatially uniform across 14 HUC12s with wet antecedent; sub-detection even at 0.45" if recession baseflow dominates. - Richland: ~0.25" with concentrated headwater bullseye on moist soil; ~0.5" otherwise; karst noise floor ±0.03–0.05 ft. - Bear Cr: dry-antecedent threshold >0.9–1.1" (Events 11–13 confirm); wet-antecedent threshold ~0.4" (Event 17). High structural absorption. - Harriet: direct-zone responses small unless concentrated >0.6"; dominated by upstream propagation.

Transfer ratios — Boxley (height ft/inch zone QPE): - Dry, low-intensity: 0.36 (Event 5, 0.55") - Dry, high-intensity (>0.5"/hr): 0.62–0.82 (Events 7, 9, 11 — intensity overruns pool storage) - Moderate: 0.56–0.67 (Events 6, 10) - Wet/moist: 0.71–0.84 (Events 2, 3) - Very wet/wet, primed: 1.30 (Event 4) — pool-fill primer mechanism - Very wet + extreme intensity: 2.28 (Event 19 Day 100) — study high

Transfer ratios — Pruitt: - Antecedent-INDEPENDENT at ~1.4 ft/in when forcing is concentrated and routing contribution is present. Dry (Event 9: 1.44) ≈ wet (Events 8, 15: 1.4–1.9) ≈ moist (Event 10: 1.6). - Distributed sub-1" forcing on dry zone with no upstream signal: <0.1 ft/in (Events 12, 17 anomalies — local-zone-only responses are sub-detection at Pruitt). - Two-mode response (Events 12, 15): surface peak at 4–7 hr lag + karst/subsurface secondary at 19–22 hr — consistent with local-knowledge karst losses between Ponca and Pruitt.

Transfer ratios — Richland: - Dry/recovering: 0.16–0.25 ft/in (Events 9, 11; Event 19 Pulse 3 ~0.24) - Dry, event-scale: 0.73 (Event 11, 1.32" zone) - Moist: 0.90 (Event 13, event-scale anchor) - Moist + high intensity: 1.34 ft/in (Event 10, 1.43" with 0.83"/hr peak) — moist + intensity can exceed wet - Wet/saturated, out-of-bank: 2.69 ft/in (Event 15, 8.89 ft peak) — channel-geometry nonlinearity - Recession-falling-limb: 0.5–0.6 ft/in (Event 8 secondary)

Transfer ratios — Bear Cr (most stage-and-antecedent-sensitive): - Dry (Event 11): 0.09–0.17 ft/in — extreme damping - Dry + intensity (Event 9): 0.35 ft/in - Moist (Events 7, 12): 0.22–0.25 ft/in - Moist + bimodal compound (Events 8, 15): 0.97–1.53 ft/in - Wet + low pre-event stage (Event 16): 2.08 ft/in — study high - Wet + mid-stage recession (Event 17): 0.51 ft/in - Bear Cr transfer is stage-AND-antecedent-dependent. Dry → ~0.1; moist → ~0.3; wet+low-stage → ~2.0; wet+mid-stage → ~0.5.

Transfer ratios — St. Joe / Harriet (cumulative event-scale, ft/in zone-avg): - St. Joe: dry 1.09 (Event 11), moist 1.3–2.1 (Events 8, 10, 19 secondary), wet ~3.0 (Event 15 flood scale). Wet-amplification factor ~3×. - Harriet: dry 0.76 (Event 11), moist ~0.8–1.4, wet ~1.4–3.0 (Event 15). Tracks St. Joe with slight attenuation.

Lag times — mainstem cascade peak-to-peak (lag = peak QPE-or-upstream-peak to peak gauge response): - Boxley→Ponca: 1.75–5 hr (1.75 hr at flood, 3–5 hr at moderate, 5 hr+ at low flow) - Ponca→Pruitt (velocity ∝ Q^0.35 confirmed across 5 events): ~5.75 hr at >1500 cfs (1.2 m/s) | 7.5 hr at ~500 cfs (0.93 m/s) | 10.75 hr at ~300 cfs (dry) | 14–15 hr at ~150–180 cfs (low flow, 0.46 m/s) | 14 hr at ~200 cfs (Event 6) - Pruitt→St. Joe: 11–14 hr wet/flood | 17–23 hr dry/low-flow | Slow segment — accounts for most of dry-condition cascade extension (~46% slower than wet) - St. Joe→Harriet: 8–11 hr — ANTECEDENT-INSENSITIVE. 6 hr at flood (Events 8, 15); 8.25 hr at moderate (Event 19); 11.25 hr at dry-mainstem (Event 11). Channel geometry between these two gauges is uniform/less karst — minimal antecedent dependence. - Total Boxley→Harriet: 27 hr at flood (Event 2) | 33 hr at wet (Events 4, 7) | 36 hr at dry (Event 11) | 44.5–52 hr at dry+low-flow (Events 9, 10 secondary).

Wet/dry asymmetries: 1. Boxley pool-fill mechanism: dry-antecedent forcing must first fill upstream pools before reaching gauge. Threshold ~0.5"; <0.5" produces sub-detection; >0.8" produces 8–11 hr delayed accelerating rise; ≥1.0" with primer or wet antecedent produces 2–4 hr rapid response. Local knowledge confirmed. 2. Pruitt's antecedent-independence at high concentrated forcing but extreme sensitivity to upstream routing presence (zero local response possible at 0.8"+ if no upstream signal). 3. Bear Cr structural damping under dry (0.09–0.17 ft/in) vs amplification under wet+low-stage (2.0+ ft/in) — 10–20× range. Stage-on-rating-curve matters as much as antecedent moisture. 4. Richland intensity-dependence: at moist antecedent, high-intensity bursts (>1"/hr) produce 4–5 hr flash with 1.3+ ft/in transfer, exceeding even saturated-wet ratios at lower intensity. 5. St. Joe distributed-forcing dilution: same total QPE produces lower per-inch transfer when spatially distributed across all 14 HUC12s vs concentrated bullseyes (Event 2 vs Event 4 vs Event 15).

Forcing-magnitude and spatial-coherence findings: - Event 12 paradox: Moist-antecedent transfer ratios LOWER than dry-antecedent Event 11 when Event 12 forcing was distributed sub-1.0" across HUC12s. Threshold effect: below ~1.0" zone-average, response efficiency drops regardless of antecedent. Initial abstraction + interception absorbs distributed small bursts even on wet soil. - Spatial coherence matters: Event 11 (broad watershed-wide pulse, 09–12 CDT cores) produced larger responses than Event 12 (spatially shifting two-day distribution with no single coherent pulse) despite Event 12's wetter antecedent and similar zone-totals.

Upstream-only flood attenuation (Event 19 finding): - Pruitt flood (2130 cfs) and Ponca flood (1920 cfs) on Day 100 attenuated through routing/channel storage to St. Joe 2900 cfs (36% of 8000 cfs flood threshold) and Harriet 2860 cfs (30% of 9370 cfs flood threshold). - Upper-basin-only flood-scale events do NOT propagate flood conditions to mid/lower mainstem. Contrast with Event 15 (watershed-wide forcing) where flood-threshold was exceeded at all four mainstem gauges with cfs and at St. Joe 1.8× / Harriet 1.6× threshold. - Implication: flood-warning logic must distinguish upper-only vs watershed-wide forcing patterns from QPE. - Recession-tail corollary (Day 102): even an attenuated upstream-only flood produces a sustained multi-day Optimal recreation window at the lower mainstem — St. Joe and Harriet remained mid-Optimal (1270/1650 cfs) ~Day 4 of the Event 19 recession. The flood pulse is brief and upstream-confined, but the elevated-baseflow tail is multi-day and basin-wide.

Signal-separation logic — validated across multiple events: - Richland gauge as St. Joe-area discriminator: Event 10 (Richland rose → St. Joe primary peak Richland-driven), Event 16 (Richland flat → St. Joe rebound Cave/Big Cr-driven), Event 19 (Richland flat → St. Joe peak pure mainstem propagation). Reliable diagnostic. - Bear Cr gauge as Harriet-area discriminator: Events 5, 16, 19 all show Bear Cr flat while Harriet rose = mainstem propagation confirmed. Event 8 shows Bear Cr surged independently while Harriet also rose = compound source. Reliable diagnostic. - St. Joe double-peak structure observable when Richland flashes earlier than Pruitt mainstem signal arrives (Event 10): primary tributary peak + secondary mainstem peak separated by ~22 hr at low-mid flow.

Recreational threshold crossings observed in study: - Hailstone Low-Floatable (Boxley 3.7 ft): crossed in Events 4 (briefly), 8 (32 hr — exceeds local "rarely >24 hr"), 19 (Day 100 5.50 ft Optimal+). - Boxley-Ponca section (3.2 ft): crossed in Events 2, 4, 8, 19. - Upper Richland Low-Floatable (Richland 3.2 ft): crossed in Events 4 (~14.5 hr), 8 (peak 4.89 ft ~Optimal), 15 (peak 8.89 ft far above High). 3 confirmed runs in study window — consistent with "3–7 runs per year" local expectation. - Mainstem Optimal durations from major events: Events 4+5+6 stacking → ~21 days continuous Optimal at St. Joe/Harriet (extended by secondary forcing); Event 15+16 stacking → ~15+ days; Event 19 → ~5+ days (confirmed: still mid-Optimal at St. Joe/Harriet on Day 102, Day 4 of recession).

Open Questions / Next Watches

1. Spatial-coherence threshold quantification: Below what zone-aggregated coherence does forcing-magnitude dominance break down? Event 12 vs Event 11 contrast is qualitative; need more data points.
2. Pruitt's local-zone forcing transfer in absence of upstream signal: Event 17 produced near-zero Pruitt response from 1.02" Cove Cr. Confirm with another upstream-dry / Pruitt-zone-wet event.
3. Bear Cr stage-dependence function: Refine the rating-curve geometry / floodplain breakout point. Current estimates: <3.5 ft pre-event stage → 1.5–2.0 ft/in; 3.5–4.0 ft → 0.5–1.0 ft/in; >5.0 ft → <0.5 ft/in (out-of-bank flattening).
4. March 2024 reference event signature: Event 15 approached but did not equal it (15.05 ft / 14,500 cfs vs ~15 ft / ~17,000 cfs). Distributed Event 15 forcing vs concentrated 2024 Richland-bullseye explains discrepancy; need a Richland-concentrated 3"+ event to fully calibrate.
5. Bear Cr long-lag character: 26 hr (Event 7) vs 9 hr (Event 9) vs 4 hr (Event 16) — spans 6× range. Lag appears antecedent-driven (dry slow, wet fast) but n=small per condition. Confirm pattern.
6. Ungauged-watershed monitoring gap: Event 15 deposited 6.087" on Long Cr (HUC 0505) with 1.584"/hr peak — flash flood potential downstream of Harriet invisible to gauge network. Recommend QPE-only watch criteria for ungauged tier.
7. St. Joe→Harriet antecedent-insensitivity: 4 events confirm 8–11 hr at high flow regardless of conditions. Investigate channel-geometry / floodplain characteristics that produce this anomaly relative to upstream segments. Day 102 adds: the two gauges also converge to near-identical absolute recession rates (~32 vs 37 cfs/hr) even at elevated flow — segment is hydraulically uniform.
8. Long-term baseflow trajectory: Day 65 was not a floor; Days 75–79, 86, 95, 97 all set new lows. Is there a true seasonal floor below current observations, or does drying continue indefinitely without recharge? Need a 30+ day dry stretch to test. (Spring 2026 has been event-rich; no such stretch yet observed post-Event-19.)
9. Pruitt karst secondary peak quantification: Two-mode response observed in Events 12, 15. Always present at lower magnitude under wet conditions, or only when specific Pruitt-zone HUCs (Cove, Hoskin) get concentrated forcing? Watch for next moderate Pruitt-zone-concentrated event.
10. Detection-threshold lower bound under saturated-recession conditions: Days 95, 97 baselines show entire watershed in clean recession; subsequent sub-threshold pulses (0.1–0.4") get absorbed completely. Establish minimum forcing required to perturb a flat-recession watershed at each gauge.
11. High-flow recession-rate calibration (new, Day 102): Day 102 provides a clean ~Day-1–2 post-peak high-flow recession-only day (St. Joe ~32 cfs/hr, Harriet ~37 cfs/hr at ~1300–2500 cfs). Combined with Days 95/97 pipeline-empty rates (~10–15 cfs/hr at low flow), this begins to define a flow-dependent recession-rate curve. Collect more post-peak recession-only days to fit dQ/dt vs Q.

End of document.