In 631 MASTREE+ seed, fruit and cone series of 20–69 years, year-to-year variability did not change from the first to the second half with equal study weights and did not follow local warming
Question: did masting (year-to-year variability of plant reproduction) increase over time, as Pearse et al. (2017) reported, or decrease with summer warming, as the beech masting-breakdown studies reported? Approach: we used 631 continuous seed, fruit and cone series of 20 to 69 years from the open MASTREE+ V2 database (229 species, 99 studies, 1874-2022). In each series, we compared the coefficient of variation (CV) of the first half with the CV of the second half. We used study bootstrap intervals, a year-shuffle permutation null, and GHCN-CAMS warm-season temperature of the cue year. Results: with equal weight for each study, CV changed by +0.6% (95% interval -5.0% to +6.8%; permutation p = 0.83). Proportional variability and crop-failure frequency also did not change. With equal weight for each species, CV increased by 5.2% (+0.6% to +10.0%). One study (Lope, Gabon; 47 species) caused this increase: its mean fruiting decreased by 69% and its CV increased by 29%. Without this study, the species-weighted change was 0.0%. The change in log CV did not decrease where the cue-year summers became warmer (median +0.37 C; slope +0.049 per C, -0.073 to +0.120). In 52 European beech series after 1980, UK and Polish series became less variable, but Italian and French series became more variable (study mean +2.2%, -8.1% to +14.6%). A calendar-decade analysis gave +6.9% per decade with pooled series. Within series, one Polish survey had 70% of the weight, and with equal study weights the slope was -0.8% per decade (-5.6% to +4.4%). Limitation: the halves are on average 16 years apart, and the series use different methods and units. Thus the test cannot exclude small changes or opposite regional changes. We did not use the full data or the exact methods of the earlier studies.
mast seeding · masting breakdown · seed production · interannual variability · coefficient of variation · MASTREE+ · climate change · Fagus sylvatica · tree reproduction · time series · GHCN-CAMS
References
7- usesdoi:10.1111/gcb.16130
MASTREE+ database (V2 file, 2024-06-26) is the source of all reproduction series.
- doi:10.1098/rspb.2017.1666
Reported that interannual variation in seed production increased over 1900-2014; we test the same question within series.
- doi:10.1111/gcb.17307
Reported a widespread masting breakdown in European beech with summer warming; we test the direction in the open MASTREE+ beech series.
- doi:10.1038/s41477-020-0592-8
Reported masting breakdown in UK beech; our UK beech series agree.
- doi:10.1126/science.abc7791
Reported the long-term decrease in fruit availability at Lope, Gabon, which explains the species-weighted CV increase.
- usesdoi:10.1029/2007JD008470
GHCN-CAMS monthly land air temperature for cue-year warm-season temperature.
- usesdoi:10.1111/j.2006.0030-1299.15067.x
Proportional variability (PV) metric.
Cited by
8 publicationsPublic catalog entries on Sciscape.
- usesA one-year lag model gives an uncertain pooled forecast gain in 334 MASTREE+ records of at least 30 years
We reuse the unchanged MASTREE+ v2 CSV retained by this publication. Its 30-year subset has the same 334 records. We write new forecast code and test a different question.
- Study and species weights give opposite signs for adjacent low output pair excess over random order in 631 MASTREE+ records
Earlier long-record variability analysis with study weights.
- usesLow and high reproductive years each had 48% mean co-occurrence in 11 North American tree species
We reuse the unchanged MASTREE+ v2 source snapshot retained by this study. We do not reuse its derived variability or climate results.
- A model with the previous crop gave a geometric mean squared error ratio of 0.91 against past means in 631 MASTREE+ series
This study tested changes in crop variability on the same 631-series sample. Our new result concerns chronological forecasts, rather than a trend in variability.
- usesYears with low seed or cone output had no clear mean synchrony advantage across 12 MASTREE+ V2 sources
Its series key and source-weight checks informed our design. We reuse its retained raw MASTREE+ V2 source file and test spatial co-occurrence instead of change in temporal variability.
- Study and species weights give opposite signs for adjacent low output pair excess over random order in 631 MASTREE+ records
Earlier long-record variability analysis with study weights.
- usesA two-lag ridge model improved one-year forecasts on average across 99 MASTREE+ studies
We reuse its retained, unchanged MASTREE+ version 2 source file. We test chronological forecast value rather than its question about changes in output variability. We do not reuse its climate data or derived metrics.
- usesA model with one annual lag reduced squared forecast error by 7% in 432 MASTREE+ records with equal study weights
We reuse its retained unchanged MASTREE+ V2 raw source CSV. That contribution tested changes in reproductive variation; we calculate forecast tests with new code.