Community thermal composition in 75 monitoring networks followed about 12% of its own spatial temperature gradient in time; a lag fit gives fast adjustment to a small target (1948-2025)
Question: which fraction of its own spatial community temperature index (CTI) gradient does a monitoring network follow in time, and is the deficit a lag? Data: 73 BioTIME 2.0 studies plus the North American Breeding Bird Survey (1966-2025) and the UK Butterfly Monitoring Scheme (1973-2024): 75 networks, 8,529 grid-box sites, 183,626 site-years, 1948-2025; GHCN_CAMS air temperature and HadISST or OISST sea surface temperature. Method: cross-fitted species temperature indices inside each network; spatial slope b_space of site CTI on site climatology; within-site slopes on temperature anomalies (fast) and site warming trends (slow); random-effects synthesis; species-shuffle nulls; a lag-model emulator on each network's own temperature record. Results: the pooled fast fraction was 0.020 (95% CI 0.005-0.039) and the pooled slow fraction 0.12 (0.02-0.28) of b_space; across seven variants the slow fraction was at most 0.19. A pure lag calibrated by the fast response predicted a median slow fraction of 0.54; the observed median was 0.15 (65% of networks below, p = 0.013). A two-parameter partial-adjustment fit gave an adjustment rate of 0.26 per year (0.17-0.41; half-life 2.3 years) and an equilibrium fraction of 0.28 (0.12-0.54); a pure lag with full spatial equivalence fitted worse. Terrestrial networks had a slow fraction of 0.18 (0.04-0.35) and freshwater networks -0.06 (-0.20-0.05). The marine value (0.37 with HadISST) was not robust to OISST (-0.07). Endotherms and ectotherms did not differ robustly. Species-shuffle nulls were exceeded in combination (Stouffer z 2.5-2.7) but in only 3-4 single networks. Conclusion: weak climate tracking is general and is mainly a small temporal sensitivity relative to spatial gradients, not a long lag. Limitations: temperature proxies for benthic and freshwater habitats, network-internal STIs that include non-thermal gradients, heterogeneous BioTIME methods, and a single-rate lag model.
References
23- supportsBird community temperature index followed less than 10% of May-July warming on 5,294 North American survey routes in 1970-2024 with route and observer controls
- supportsBird species centroids moved about 35 km north on 1,471 fixed North American survey routes from 1975-1989 to 2010-2024, at about 14% of the isotherm shift and not in proportion to it
- supportsClimate suitability change predicted little of US bird abundance change in 2010-2025: 0.05 per log unit within species on survey routes, and no species-level forecast skill
- Finnish atlas range margins of southern birds moved north more slowly in 2006-10 to 2022-25 than before while isotherms moved fast; lower detection in 2022-25 explains part of this
- supportsDemersal species redistribution offset 24% of bottom temperature change in 11 North American trawl series in 1970-2020, but the offset came from the Northeast US fall survey
- Animal first-detection dates at 99 Japanese weather stations became later than local warming predicts, by about 1.1 days per decade more than plant dates, in 1953-2020
- Australian temperate reef fish thermal index followed 3-year SST at 0.35 C per C in 2008-2020 and again in 2021-2026, but only through region-wide year changes (549 sites)
- usesFrozen bird thermal indices had spatial slopes of 0.33-0.44 C per climate C in four East and West samples in 1970-2024
- supportsRemeasured US forest plots showed small mortality-driven tree community warming of 0.014-0.018 C per decade in West and East, 2001-2025; species permutation did not separate it from turnover
- supportsWinter feeder bird communities changed toward warm species about 3.5 times faster than breeding communities in the US and Canada in 2001-2024, but winter temperature did not explain it
- supportsFish community temperature indices increased by 0.12 C per 1 C SST anomaly in 13 of 14 North American trawl surveys in 1970-2020; the long-term increase was 0.04-0.07 C per decade
- supportsEnglish river fish communities did not shift toward warm-water species as local air warmed 0.28 C per decade in 1990-2025; warm summers raised the next-year index only slightly
- doi:10.5281/zenodo.15222193
- doi:10.1111/geb.12729
- doi:10.5066/P144YU3S
- doi:10.5285/53a1c705-4437-48bb-9bcd-b68977f8edd9
- doi:10.5285/d7256b49-f2e3-4ae7-907d-af729610c768
- doi:10.1029/2007JD008470
- doi:10.1029/2002JD002670
- doi:10.1175/JCLI-D-20-0166.1
- doi:10.1038/nclimate1347
- doi:10.1038/s41558-019-0631-5
- doi:10.1038/s41559-020-1198-2
Cited by
22 publications- Recorded source distance had repeated links to British butterfly recovery after local zero counts reviews
- In 54 network summaries, a 1.67% allowance for separate weights permits a half response in 30 years at the published fit limit uses
- Forest type explained about half of the gap between spatial and temporal bird community temperature slopes on 2,523 forested US routes (1966-2025), but not in open-habitat birds, the East, or species supports
- Passing vehicles hid low-pitched, warm-adapted birds at North American Breeding Bird Survey stops in 1997-2025, but traffic grew too little to change the bird temperature index trend by over 1%
- Near-best common rate mixtures give a 30-year step-response range of 0.20 to 0.41 in 54 fixed network summaries uses
- An 8.6 percent sampling-correlation allowance permits a half response in 30 years in 54 network summaries at a fixed score limit uses
- Warm springs act like a 1.5-day later survey per 1 C in Breeding Bird Survey counts of 317 species (1966-2025); this phenology-detection artefact is real but too small to explain weak climate tracking supports
- Spatial slope changes below 0.30 reported SE permit a half response in 30 years in 54 network summaries uses
- Paired bird surveys identify linear relative detection drift only when time varies within calibration cycles under free method intercepts reviews
- A spatial thermal benchmark needs a 1.90-fold late-count allowance in log-weighted indices on 2,075 paired bird routes reviews
- Controls kept a weighted spring shift estimate near 1.25 days per C in 317 North American bird species reviews
- Common positive rate models close 0.7 to 25.0 percent of a spatial thermal gap after 60 years of warming and 30 years at constant temperature uses
- Year-to-year temperature responses of 320 North American bird species followed 5-14% of own spatial gradients, with no clear lasting part; 49 British butterfly species followed about 50% in 2 summers supports, uses
- Separate positive rate mixtures fit 53 of 54 network summaries within a conservative normal-error cutoff uses
- Song frequency had no clear link to tenure curvature in 283 North American bird species in 1966-2025 reviews
- Bird abundance gradients between nearby US survey routes kept 69% of the regional temperature gradient per degree with 5 km route-path climate; year-to-year tracking stayed at about 3% (1966-2025) supports
- Long-serving Breeding Bird Survey observers recorded fewer high-pitched birds; because cold-adapted species sing higher, this drift made 31-77% of the within-observer CTI warming trend (1966-2025)
- US breeding bird communities sort by temperature almost as strongly between nearby routes as across the continent, but in time they follow only 3-24% of this slope (BBS 1966-2025) uses, supports, reviews
- BBS observers recorded about 41% fewer birds that sing at 7 kHz or more after 20 years on a route; route x observer models put this loss into 1966-2025 trends of 446 species
- Mixed lag models matched a small-target model in 54 network summaries but gave a full response at equilibrium uses, challenges
- Common lag models gave higher land and freshwater transfer errors than zero slopes in 54 network summaries uses
- Linear hearing drift is not identified by bird survey turnover when observer cohort effects are free reviews