Blossom Frost Risk: Orchard Pollination Window

Winter & Spring Climate

Winter severity 1.0×

Spring warming rate 1.0×

Day-to-day climate variability 2.0 °C

Live Readouts

Chill requirement met

—

Bloom starts (chill + GDD met)

—

Bloom window

—

Successful pollination days

—

Frost-damaged bloom days

—

Too-cold-for-bees days

—

Calendar Legend

Pollinated (bees flying, no frost) Frost-damaged (below frost threshold) Too cold for bee flight Chill accumulation (dormant) GDD accumulation (pre-bloom)

How it works

Bloom timing is a genuine two-stage accumulator run over a simulated year of daily temperatures, not a fixed calendar date:

chill(d)  += 24h  if 0°C ≤ T(d) ≤ 7.2°C     (winter, from day 1)
  → chillDay: first day cumulative chill ≥ 1500h (endodormancy broken)
GDD(d)    += max(0, T(d) − 4.5°C)            (from chillDay+1)
  → bloomStart: first day cumulative GDD ≥ 55
bloom window = bloomStart … bloomStart+15  (16 days)

for each day d in the bloom window:
  frost = dayMin(d) < −1.5°C   → blossoms damaged that day
  bees  = dayMax(d) > 13°C     → bees can fly & pollinate that day
  pollinated(d) = bees AND NOT frost

Winter severity — scales the seasonal temperature swing; a colder winter still meets the chill requirement (it's easy to satisfy near 0–7°C), but the same cold nights persist later into the accumulation.
Spring warming rate — scales how fast temperatures rise once the chill requirement is met; a faster spring reaches the GDD threshold sooner, pulling bloom earlier into the season when nights are still frost-prone.
Climate variability — adds day-to-day noise (an autocorrelated random walk) on top of the smooth seasonal curve. It never changes bloom timing on average, but it does push individual bloom-window nights below the frost line and individual afternoons below the bee-flight line — so higher variability measurably reduces the count of days on which both conditions hold at once.
New season — redraws the whole year's daily temperature trace from a fresh random seed, re-running chill → GDD → bloom-window classification from scratch.
Run variability sweep — replays the same season dozens of times at each variability level and plots the average successful-pollination-day count, so the variability → pollination relationship is measured directly rather than only asserted.

Real-world relevance: this is the same two-stage chill/GDD phenology model orchardists and agronomists use to forecast bloom dates, and the same reason a mild-but-erratic spring can cost more fruit set than a cold-but-stable one — narrow safe-pollination windows are eroded by variance, not just by average warmth.

Variability sweep — avg. successful pollination days

Running…