Tree phenology modelling in the boreal and temperate climate zones : Timing of spring and autumn events
(2014)- Abstract
- Plant phenology in the boreal and temperate climate zones is synchronised with seasonal changes in temperature and photoperiod. For deciduous trees, timing of budburst and leaf colouring define the growing season length and express adaptations to trade-offs in growth and risks. An extended growing season due to climate warming will likely increase forest productivity, however these potential benefits may be outweighed by an increase in risks such as frost damage.
In order to estimate the impact of climate change on forest phenology, the reliability of phenology models needs to be assessed. In this thesis, the ability of phenology models to capture inter-annual and spatial variation in budburst and leaf colouring were evaluated for... (More) - Plant phenology in the boreal and temperate climate zones is synchronised with seasonal changes in temperature and photoperiod. For deciduous trees, timing of budburst and leaf colouring define the growing season length and express adaptations to trade-offs in growth and risks. An extended growing season due to climate warming will likely increase forest productivity, however these potential benefits may be outweighed by an increase in risks such as frost damage.
In order to estimate the impact of climate change on forest phenology, the reliability of phenology models needs to be assessed. In this thesis, the ability of phenology models to capture inter-annual and spatial variation in budburst and leaf colouring were evaluated for five tree species that are important in Swedish forest management; birch, beech, oak, Norway spruce and Scots pine. Model simulations were assessed in relation to model structure, the models representation of tree physiology processes and the calibration -, temperature - and phenology data used. The novelty of the thesis is that simulations were carried out across large regions using extensive phenological datasets which consist of observations of more than 1000 trees, with up to 60 observation-years per tree.
The results indicate that the model structure influence the models sensitivity to calibration data and to temperature conditions. Spatial differences in trees response to environmental cues is not well represented in the models. The models do not consider differences in provenance-specific requirements and therefore was the accuracy of the simulations influenced by the models being tuned to the average phenological response of the calibration data. Overall, the budburst models overestimated the temperature effect. In colder regions where budburst usually occurs later, budburst was simulated to occur too late, and in warmer regions where budburst occurs earlier, budburst was simulated to occur too early. The more accurate budburst models were in general structurally simple. They considered the effect of warm spring temperatures, thereby assuming full dormancy release without considering winter conditions. Leaf colouring was better estimated by average day of leaf colouring than by the model simulations.
In conclusion, the results indicate that the models do not capture the phenology across large regions well enough to be considered reliable for climate change assessments, emphasising the lack of mechanistic understanding of phenological processes. (Less) - Abstract (Swedish)
- Popular Abstract in Swedish
I den boreala och tempererade klimatzonen är växternas fenologi synkroniserad med årstidsväxlingar i temperatur och dagslängd. För lövträd påverkar tidpunkten för knoppsprickning och lövfärgning både tillväxt och risktagande. En förlängd växtsäsong till följd av ett varmare klimat innebär sannolikt högre produktivitet inom skogsbruket, men kan också leda till ökad risk för exempelvis frostskador.
För att uppskatta klimatförändringarnas effekt på skogens fenologi behövs tillförlitliga fenologimodeller. I den här avhandlingen utvärderades hur väl fenologimodeller fångar mellanårs och rumslig variation i tidpunkt för knoppsprickning och lövfärgning för fem trädarter viktiga inom svensk... (More) - Popular Abstract in Swedish
I den boreala och tempererade klimatzonen är växternas fenologi synkroniserad med årstidsväxlingar i temperatur och dagslängd. För lövträd påverkar tidpunkten för knoppsprickning och lövfärgning både tillväxt och risktagande. En förlängd växtsäsong till följd av ett varmare klimat innebär sannolikt högre produktivitet inom skogsbruket, men kan också leda till ökad risk för exempelvis frostskador.
För att uppskatta klimatförändringarnas effekt på skogens fenologi behövs tillförlitliga fenologimodeller. I den här avhandlingen utvärderades hur väl fenologimodeller fångar mellanårs och rumslig variation i tidpunkt för knoppsprickning och lövfärgning för fem trädarter viktiga inom svensk skogsbruk; björk, bok, ek, gran och tall. Modellsimuleringarna analyserades i relation till modellkonstruktion, modellernas representation av trädens fysiologiska processer och till den kalibrerings-, temperatur- och fenologidata som användes. För att kunna göra en bra utvärdering användes observationer från fler än 1000 träd runt om i Europa, med upp till 60 års observationer per träd.
Resultaten indikerar att modellernas konstruktion påverkar modellerna känslighet för kalibreringsdata och temperaturförhållanden. Modellerna lyckades inte fånga rumsliga skillnader i trädens respons till temperatur och dagslängd. Eftersom modellerna inte tar hänsyn till att träd med olika härkomst kan ha olika ljus- och temperaturbehov, påverkades precisionen i simuleringarna av modellkalibreringen genom att modellerna anpassats till den genomsnittliga fenologiska responsen i kalibreringsdata. Överlag överskattade knoppsprickningsmodellerna temperatur-effekten. I kallare regioner där knoppsprickning sker relativt sent simulerades knoppsprickning för sent och i varmare regioner där knoppsprickning sker tidigare simulerades knoppsprickning för tidigt. De bättre knoppsprickningsmodellerna var relativt enkla i sin konstruktion och inkluderade effekten av varma vårdagar medan vintervilan antogs var helt bruten genom att inte inkludera vinterförhållanden. Tidpunkt för lövfärgning uppskattades bättre av medeldag än med modellsimuleringar.
Sammanfattningsvis indikerar resultaten att modellerna inte simulerar fenologi tillräckligt bra för att kunna anses vara tillförlitliga i klimatförändringsstudier, vilket betonar bristen på mekanistisk förståelse av fenologiska processer. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/4698893
- author
- Olsson, Cecilia LU
- supervisor
- opponent
-
- Richardson, Andrew, Dept of Organismic and Evolutionary Biology, Harvard University
- organization
- publishing date
- 2014
- type
- Thesis
- publication status
- published
- subject
- pages
- 145 pages
- publisher
- Department of Physical Geography and Ecosystem Science, Lund University
- defense location
- Pangea auditorium
- defense date
- 2014-11-14 10:00:00
- ISBN
- 978-91-85793-43-3
- language
- English
- LU publication?
- yes
- id
- 119e44db-7e5b-4643-b896-62388890feb3 (old id 4698893)
- date added to LUP
- 2016-04-04 11:40:03
- date last changed
- 2019-03-22 11:38:54
@phdthesis{119e44db-7e5b-4643-b896-62388890feb3,
abstract = {{Plant phenology in the boreal and temperate climate zones is synchronised with seasonal changes in temperature and photoperiod. For deciduous trees, timing of budburst and leaf colouring define the growing season length and express adaptations to trade-offs in growth and risks. An extended growing season due to climate warming will likely increase forest productivity, however these potential benefits may be outweighed by an increase in risks such as frost damage.<br/><br>
In order to estimate the impact of climate change on forest phenology, the reliability of phenology models needs to be assessed. In this thesis, the ability of phenology models to capture inter-annual and spatial variation in budburst and leaf colouring were evaluated for five tree species that are important in Swedish forest management; birch, beech, oak, Norway spruce and Scots pine. Model simulations were assessed in relation to model structure, the models representation of tree physiology processes and the calibration -, temperature - and phenology data used. The novelty of the thesis is that simulations were carried out across large regions using extensive phenological datasets which consist of observations of more than 1000 trees, with up to 60 observation-years per tree. <br/><br>
The results indicate that the model structure influence the models sensitivity to calibration data and to temperature conditions. Spatial differences in trees response to environmental cues is not well represented in the models. The models do not consider differences in provenance-specific requirements and therefore was the accuracy of the simulations influenced by the models being tuned to the average phenological response of the calibration data. Overall, the budburst models overestimated the temperature effect. In colder regions where budburst usually occurs later, budburst was simulated to occur too late, and in warmer regions where budburst occurs earlier, budburst was simulated to occur too early. The more accurate budburst models were in general structurally simple. They considered the effect of warm spring temperatures, thereby assuming full dormancy release without considering winter conditions. Leaf colouring was better estimated by average day of leaf colouring than by the model simulations.<br/><br>
In conclusion, the results indicate that the models do not capture the phenology across large regions well enough to be considered reliable for climate change assessments, emphasising the lack of mechanistic understanding of phenological processes.}},
author = {{Olsson, Cecilia}},
isbn = {{978-91-85793-43-3}},
language = {{eng}},
publisher = {{Department of Physical Geography and Ecosystem Science, Lund University}},
school = {{Lund University}},
title = {{Tree phenology modelling in the boreal and temperate climate zones : Timing of spring and autumn events}},
url = {{https://lup.lub.lu.se/search/files/5827081/4699512.pdf}},
year = {{2014}},
}