1.
Chmielewski, Frank-M.; Blümel, Klaus; Scherbaum-Heberer, Carina; Koppmann-Rumpf, Bettina; Schmidt, Karl-Heinz
In: International Journal of Biometeorology, Bd. 57, S. 287-297, 2013.
Abstract | Links | BibTeX | Schlagwörter: Climate change, Egg laying of Great Tit, Model comparison, Phenological modelling, Photoperiod, Projections of egg laying
@article{Chmielewski2013,
title = {A model approach to project the start of egg laying of Great Tit (Parus major L.) in response to climate change},
author = {Frank-M. Chmielewski and Klaus Bl\"{u}mel and Carina Scherbaum-Heberer and Bettina Koppmann-Rumpf and Karl-Heinz Schmidt},
doi = {10.1007/s00484-012-0553-7},
year = {2013},
date = {2013-01-01},
urldate = {2013-01-01},
journal = {International Journal of Biometeorology},
volume = {57},
pages = {287-297},
abstract = {The aim of this study was to select a phenological model that is able to calculate the beginning of egg laying of Great Tit (Parus major) for both current and future climate conditions. Four models (M1\textendashM4) were optimised on long-term phenological observations from the Ecological Research Centre Schl\"{u}chtern (Hessen/Germany). Model M1 was a common thermal time model that accumulates growing degree days (GDD) on an optimised starting date t 1. Since egg laying of Great Tit is influenced not only by air temperature but also by photoperiod, model M1 was extended by a daylength term to give M2. The other two models, M3 and M4, correspond to M1 and M2, but t 1 was intentionally set to 1 January, in order to consider already rising temperatures at the beginning of the year. A comparison of the four models led to following results: model M1 had a relatively high root mean square error at verification (RMSEver) of more than 4 days and can be used only to calculate the start of egg laying for current climate conditions because of the relatively late starting date for GDD calculation. The model failed completely if the starting date was set to 1 January (M3). Consideration of a daylength term in models M2 and M4 improved the performance of both models strongly (RMSEver of only 3 days or less), increased the credibility of parameter estimation, and was a precondition to calculate reliable projections in the timing of egg laying in birds for the future. These results confirm that the start of egg laying of Great Tit is influenced not only by air temperature, but also by photoperiod. Although models M2 and M4 both provide comparably good results for current climate conditions, we recommend model M4\textendashwith a starting date of temperature accumulation on 1 January\textendashfor calculating possible future shifts in the commencement of egg laying. Our regional projections in the start of egg laying, based on five regional climate models (RCMs: REMO-UBA, ECHAM5-CLM, HadCM3-CLM, WETTREG-0, WETTREG-1, GHG emission scenario A1B), indicate that in the near future (2011\textendash2040) no significant change will take place. However, in the mid- (2041\textendash2070) and long-term (2071\textendash2100) range the beginning of egg laying could be advanced significantly by up to 11 days on average of all five RCMs. This result corresponds to the already observed shift in the timing of egg laying by about 1 week, due mainly to an abrupt increase in air temperature at the end of the 1980s by 1.2 K between April and May. The use of five regional climate scenarios additionally allowed to estimate uncertainties among the RCMs.},
keywords = {Climate change, Egg laying of Great Tit, Model comparison, Phenological modelling, Photoperiod, Projections of egg laying},
pubstate = {published},
tppubtype = {article}
}
The aim of this study was to select a phenological model that is able to calculate the beginning of egg laying of Great Tit (Parus major) for both current and future climate conditions. Four models (M1–M4) were optimised on long-term phenological observations from the Ecological Research Centre Schlüchtern (Hessen/Germany). Model M1 was a common thermal time model that accumulates growing degree days (GDD) on an optimised starting date t 1. Since egg laying of Great Tit is influenced not only by air temperature but also by photoperiod, model M1 was extended by a daylength term to give M2. The other two models, M3 and M4, correspond to M1 and M2, but t 1 was intentionally set to 1 January, in order to consider already rising temperatures at the beginning of the year. A comparison of the four models led to following results: model M1 had a relatively high root mean square error at verification (RMSEver) of more than 4 days and can be used only to calculate the start of egg laying for current climate conditions because of the relatively late starting date for GDD calculation. The model failed completely if the starting date was set to 1 January (M3). Consideration of a daylength term in models M2 and M4 improved the performance of both models strongly (RMSEver of only 3 days or less), increased the credibility of parameter estimation, and was a precondition to calculate reliable projections in the timing of egg laying in birds for the future. These results confirm that the start of egg laying of Great Tit is influenced not only by air temperature, but also by photoperiod. Although models M2 and M4 both provide comparably good results for current climate conditions, we recommend model M4–with a starting date of temperature accumulation on 1 January–for calculating possible future shifts in the commencement of egg laying. Our regional projections in the start of egg laying, based on five regional climate models (RCMs: REMO-UBA, ECHAM5-CLM, HadCM3-CLM, WETTREG-0, WETTREG-1, GHG emission scenario A1B), indicate that in the near future (2011–2040) no significant change will take place. However, in the mid- (2041–2070) and long-term (2071–2100) range the beginning of egg laying could be advanced significantly by up to 11 days on average of all five RCMs. This result corresponds to the already observed shift in the timing of egg laying by about 1 week, due mainly to an abrupt increase in air temperature at the end of the 1980s by 1.2 K between April and May. The use of five regional climate scenarios additionally allowed to estimate uncertainties among the RCMs.