Effect of climate warming to the dormancy and dehardening of pedunculate oak (Quercus robur L.) seedlings in assumed climate conditions of the year 2030 and 2100

UNIVERSIDAD POLITECNICA DE VALENCIA E S C U E L A P O L I T E C N I C A S U P E R I O R D E G A N D I A L i c e n c i a d o e n C i e n c i a s A m b i e n t a l e s “Effect of climate warming to the dormancy and dehardening of pedunculate oak (Quercus robur L.) seedlings in assumed climate conditions of the year 2030 and 2100” TRABAJO FINAL DE CARRERA Autor/es: Rubén Oliver Tomàs Director/es: Dr. Pertti Pulkkine Olga Mayoral García-Berlanga GANDIA, 2011 Environmental Sciences - L. Ciencias Ambientales Master Thesis – Trabajo Final de Carrera 05/2011 Universitat Politècnica de València University of Helsinki METLA – Finnish Forest Research Institute Rubén Oliver Tomàs Effect of climate warming to the dormancy and dehardening of pedunculate oak (Quercus robur L.) seedlings in assumed climate conditions of the year 2030 and 2100 Abstract and Acknowledgement 2 ABSTRACT The climate warming effect on the development of the pedunculate oak (Quercus robur L.) during the dormancy state and dehardening development was studied by the simulation of different growth conditions. The seed material used was obtained from 9 different origins from Poland to South Finland, to ascertain the most appropriate origin to grow in South Finland. In greenhouses they were simulated two future climatic conditions belonging to the years 2030 and 2100, using the scenario A1B, as well as two conditions of humidity. To find out the state of dormancy and the dehardening ability, we used a freezing test to simulate the injuries caused by the frosts. In dormancy experiment, the seedlings of the year 2030 presented null injuries and a high rate of survival. On the contrary, by the year 2100 the injuries were 50%. In the dehardening experiment, it was observed that the late frost was a huge threat to the survival, since the mortality increased by 30 % in two weeks. Among all the origins, the one that presents better adaptability was the one from Ruissalo. The results of this thesis supports the theory that climate warming will increase the risk of frost damage of the one year old seedlings of pedunculate oak in South Finland. Table of Contents 3 TABLE OF CONTENTS ABSTRACT TABLE OF CONTENTS 1. INTRODUCTION 1.1 Climate change 1.2 Pedunculate oak (Quercus robur L.) 1.2.1 Distribution 1.2.2 Description and morphology 1.2.3 Role in biodiversity and utilization 1.3 Plant adaptation to northern areas 1.3.1 Cold acclimation 1.3.2 Frost damage and risk of frost damage 1.4 Freeze testing 1.5 Aim of the study 2. MATERIAL AND METHODS 2.1 Material 2.2 Condition 2.2.1 Previous summer condition 2.2.2 Growth conditions during dormancy experiment 2.2.3 Growth conditions during dehardening experiment 2.3 Freeze testing 2.4 Data processing 2.5 Statistical analysis Table of Contents 4 3. RESULTS 3.1 Survival after growth period 3.2 Height before freezing test 3.3 Dormancy 3.3.1 Injuries 3.3.2 Survival 3.4 Dehardening 3.4.1 Injuries 3.4.2 Survival 3.5 The effect of seedlings size on the injuries 4. DISCUSSION 4.1 Experimental process 4.2 Effect of growth condition on oak seedlings development 4.2.1 Survival after growth period 4.2.2 Height before freezing test 4.3 Dormancy state 4.4 Dehardening development 5. CONCLUSIONS ACKNOWLEDGEMENTS 6. REFERENCES Introduction 5 1. INTRODUCTION 1.1 Climate change Some observations describing the overall image of a world warming up and other changes in the climate system are listed. The average global temperature at the surface has increased by 0.6 ± 0.2 ° C since the late XIX century. It is very likely that precipitation has increased by 0.5 to 1% per decade in the XX century at most middle and high latitudes of the Northern Hemisphere continents (IPCC 2001). Most of the observed global warming over the last 50 years is likely to have been caused by an increasing concentration of greenhouse gases, but we cannot exclude the natural climate variability in a regional scale as a potential explanation for the trend in Finland (Jylhä et al. 2004). Climate change in Intergovernmental Panel on Climate Change (IPCC 2001) usage refers to any change in climate over time, whether due to natural variability or as a result of human activity. This usage differs from that in the Framework Convention on Climate Change (UNFCCC 1999), where climate change refers to a change of climate that is attributed directly or indirectly to human activity that alters the composition of the global atmosphere and that is in addition to natural climate variability observed over comparable time periods. The most remarkable forcing agents that cause climate change have been the increases in atmospheric concentrations of greenhouse gases. The current data sets show the human influence on atmospheric concentrations of these agents, although most of the gases originate from both natural and anthropogenic. The main gases responsible for climate change are carbon dioxide (CO2), methane (CH4), nitrous oxide (N2O), and halocarbons (CFC) (IPCC 2001). The gas most responsible for climate change is also the greenhouse gas mostly produced by humans, which is the carbon dioxide. The concentration of CO 2 in the atmosphere has increased from 280 ppm in 1750 to 367 ppm in 1999. The present CO 2 concentration has not been surpassed in the last 420,000 years and has not probably occurred, even during the past 20 million years, either (IPCC 2001). From studies of climate change impacts and adaptations, Jylhä and Laapas (2009) illustrate the estimated temporal evolution of CO 2 emissions and the atmospheric CO 2 concentration in Finland (Figure 1). The tendencies are based in 3 different scenarios, which are documented by Jylhä et al. (cited in Jylhä and Laapas 2009). In our experiment we used the A1B scenario for the simulations of climate change. Introduction 6 Figure1. (a): Estimated temporal evolution of the emissions and (b): atmospheric concentration of carbon dioxide under three Special Reports on Emissions Scenarios (SRES) greenhouse gas. Figures for the SRES scenarios are based on IPCC (2001). (Jylhä and Laapas 2009 and Jylhä et al. 2009). Jylhä et al. (2004) reported that the annual mean temperatures in Finland increased by about 0.7ºC from 1901 to 2000. In contrast to temperature, the annual mean fluctuation of precipitation anomalies do not show any statistically significant trend over the 20 th century. The Finnish Research Programme of Climate Change (SILMU) proposed a climate scenario for the Nordic countries, when the scenario specifies a warming rate of 0.45ºC per decade in Finland. Johannesson (1995) suggested that temperature changes in the winter were larger (0.6ºC per decade) than in the summer period (0.3ºC per decade). Jylhä, et al. (2009) illustrates on Figure 2 the predictions of mean temperature in Finland using three scenarios. The scenario used in our experiment was the A1B. For the year 2030 the annual average temperature predicted is practically the same in all of the scenarios. However, the simulations done for the year 2100 have differences among themselves, with our scenario in the middle. The annual average is close to 5 positive degrees; for the summer period the temperature is less than in the winter, being the temperature simulated for our experiment of 4 positive degrees. It is projected that there will be an increase in global average water vapor, evaporation and precipitation. Thus, if it is assumed that the rate of relative precipitation increase is proportional to the rate of warming, it will arrive at a ratio of about 3 to 4% per ºC of warming using relatively low values of the ratio of precipitation (Johannesson 1995). Introduction 7 Figure2. Annual mean temperature in Finland modeled and projected employing three SRES scenarios. The figure is based on figures of Jylhä et al. 2009. Talkkari (1996) predicted that the most drastic changes will take place in northern latitudes, which are mostly covered by boreal forests, and the predicted global climate change is among the major factors affecting future forest development in the boreal zone. Introduction 8 1.2 Pedunculate oak (Quercus robur L.) 1.2.1 Distribution Pedunculate oak have the widest geographic distribution among European white oaks (Figure 3a). It is distributed in Europe from northern Spain to southern Scandinavia and from Ireland to Eastern Europe and reaches the Ural Mountains (Zanetto et al. 1994; Ducousso and Bordacs 2004). Most of the present South Finland was submerged when the ice receded at the end of the last ice age about 10,000 years ago. This area has been, and still is, subject to rapid land uplift. Colonization of the newly emerged ground in the South and simultaneous decline in the North have made the whole distribution of oak in Finland shift towards South and West (Figure 3). In a way, the process could be seen as the last phase of the post-glacial colonization of Europe by oaks, only now towards south (Vakkari et al. 2005). Repo et al. (2007) show that the population of pedunculate oak in Finland is strongly fragmented and growing at the northern margin of the species’ European distribution and it can only be found in the south of Finland (Figure 3b). It is the only oak species in Finland, even the closely related sessile oak (Quercus. petraea) is absent and hence introgression between the species does not affect genetic variability (Mattila et al. 1994; Vakkari et al. 2005). The cooling climate and expansion of Norway spruce led to a long-lasting decline of oak populations and, starting about 300 years ago, the stands were further fragmented and potentially restricting the flow of genes among local stands owing to increasing human pressure for more agricultural land (Mattila et al. 1994). Oak populations have also been exploited as a source of valuable timber and even destroyed on some occasions. Vakkari et al. (2005) found genetic structural differences in the populations in Finland, because of the fragmentation, and it is affected by population size and age of the population site. Planting of oak is not common in Finland, and the low number of haplotypes and the sharpness of the contact zone support the idea that manmediated seed transfer has not been very extensive in the past, either. Introduction 15 southern Finland. Repo et al. (2007) hypothesized that short growing seasons and insufficient cold hardiness during autumn, and the consequent frost damage, are key factors that restrict the northward growth of oak. 1.4. How to test cold acclimation? Due to the importance of cold acclimation for winter survival of northern plants, it was necessary to establish a method to determine levels of plant cold hardiness, dormancy and dehardening and thereby to obtain data of injury and survival levels (Lindén 2002). To standardize the freezing test protocol, Levitt (cited in Lindén 2002) proposed the following steps as basic requirements: 1. The plants must be inoculated to ensure freezing. 2. Cooling must occur at a standard rate. 3. A single freeze must be used for a standard length of time. 4. Thawing must occur at a standard rate of warming. 5. Post-thawing conditions must be standardized. Using the methodology of cooling simulation, the temperature is lowered at a rate of 1 to 2ºC/h to imitate natural frost events. A common approach is to use a fixed rate of 2 to 6ºC/h (Lindén 2002), but in nature the rate is usually less (Levitt 1972). As shown by Harrison et al. (cited in Lindén 2002), the rate of cooling is critical in the temperature range where the amount of unfrozen water is sufficient to cause injury. Lennartsson (2003) exposed that the most crucial aspects of freeze tests are the control of initiate ice formation at -2 °C to avoid supercooling, to apply slow cooling (3 °C/h) to ensure equilibrium freezing, and finally to apply slow thawing. Freezing tests can be evaluated by assessing the plant’s capacity for regrowth after freezing and the extent of visible injuries and the mortality of seedlings (Lennartsson 2003). Lindén (2002) in her doctoral thesis, Measuring cold hardiness in woody plants, lump together different alternative methods with in regard to the common controlled freezing tests. The content of dry matter or soluble carbohydrates, as well as carbohydrate composition, is often correlated with plant or tissue cold hardiness. Others investigations have show the relationship between the quality or quantity of amino acids, proteins and lipids and the level of cold hardiness (e.g. Yoshida, Khanizadeh et al., Arora and Wisniewski , Arora et al., cited in Lindén 2002). Additional Introduction 16 methods used to determine cold hardiness without freezing include electrical impedance analysis (Coleman, Repo et al., Repo et al., Väinölä and Repo, cited in Lindén 2002) and measurement of the ability to withstand plasmolysis (Siminovitch and Briggs, Levitt, cited in Lindén 2002). The indirect hardiness indicators often work well for some plants or tissues, or under certain conditions, yet none of them can be trusted as a general measure of cold hardiness in all plants (Lindén 2002). 1.5 Aim of study The main aim of this master thesis was to assess the effects of climate warming in the process of cold acclimation of one year old seedling of pedunculate oak (Quercus robur L.). The target of this study was to deal with the state of deep dormancy and the process of dehardening. In the experiment we tried to find out the difference of behavior among all origins and between Finland and more Southern seedlings. And get the most adequate origin for growth in South Finland in future conditions. We intended to find out how frost damage affects on the seedlings, analyzing the injuries and survival. To do that, the seedlings were subjected to different previous summer conditions of temperature and humidity, for the years of 2030 and 2100 simulated. All this was replicated with the aim of knowing the cold acclimation rate too. With all this, we tried to ascertain the viability of the natural oak stands in South Finland, as well as their use as ornamental trees in future years. Material and Methods 17 2. MATERIAL AND METHODS 2.1. Plant material The oak acorns (Quercus robur L.) were collected from 9 different areas (Figure 4), using seeds from the forest, orchard and urban parks (Table 1). The most seeds origin came from the southern Finland, but there was also from other countries of north of Europe, so we made two big useful groups to analyze. The first group appointed “Finland” with seeds only from the south of Finland, and the other appointed “more Southern” with seeds from southern Sweden, Latvia and north Poland. There were two different types of stands; one a natural stands origin and the other planted. In Malmi origin seed is probably source from Estonia, and the Annalapuisto seeds came from Finland (Pulkkinen pers. comm.). Thus entries has been divided into three categories, depending upon the typology of stand in which acorns were extract - Park: The stand is definitely planted and the genetic origin of the trees is undefined. - Forest stand: In all probability its origin is natural, but due to the uncertainty of its origin cannot be excluded that have been artificially planted. - Seed orchard: The trees were planted by man with the purpose to produce acorns for commercial exploitation. The material was collected during the summer of 2008 except the seeds from Poland that had been collected during the summer of 2007. Material and Methods 18 1Helsinki: - Malmi - Annalanpuisto 2Turku area: - Kaarina - Ruissalo - Paraninen 3Raasepori 4Bjuv 5Jankaulsnava 6Kozienice Figure 4. Location of the different origins of pedunculate oak acorns in Europe and Finland. Helsinki and Turku area have more than one sample used for the experiment. 1 4 3 2 6 5 Material and Methods 19 Table 1. Origin and characteristics of the oak seeds used. Group Country Origin Area Type of place Type Latitude (N) Longitude(E) Day degrees Finland Finland Helsinki Malmi Park Planted 60°14'195'' 25°1'647'' 1250 Finland Helsinki Annalanpuisto Park Planted 60°12'745'' 24°58'459'' 1250 Finland Kaarina Katariinanlaakso Forest stand Natural 60°24'713'' 22°16'344'' 1250 Finland Turku Ruissalo Forest stand Natural 60°25'525'' 22°7'512'' 1250 Finland Raasepori Tenhola Forest stand Natural 60°0' 472'' 23°5'729'' 1250 Finland Parainen Parainen Forest stand Natural 60°14'844'' 22°13'276'' 1250 More Southern Sweden Söderåsen Söderåsen Forest stand Natural 56° 5' 037" 13º 14' 574" 1550 Latvia Jankaulsnava Jankaulsnava Seed orchard Planted 56°41'187" 25°58'129'' 1650 Poland Kozienice Kozienice Forest stand Natural 51°58'553'' 21°55'117'' 1900 Material and Methods 20 2.2. Conditions 2.2.1 Previous summer condition The present study was carried out in Finnish Forest Research Institute’s Haapastensyrjäa Unit (60º 37’ 04’’, 24º 25’ 4’’, 125 meters a. s. l.) There were two dates of sowing; the first acorns were sowed the 8.04.2009 for the condition of year 2100 and the second date was the 29.04.2009 for the condition of year 2030. The seeds were sowed in a box of 40 spaces and the boxes were collocated randomly in tables of 36 boxes. The growth substrate used was 100% of peat the type of Kekkilä White 420 W with ph of 5.5 and 2.0 mS/cm of conductivity. The fertilizing used was NPK 16-4-17 with the nutrient proportion of 9.0% for NO3-N and 6.5% for NH4-N. For the same experiment (for both, dormancy to dehardening) have simulated different growing conditions in greenhouses. On the one hand we have two types of controlled temperatures that simulate the future growth condition of the Year 2100 and Year 2030. The temperature used to simulate the thermal conditions provided for the years 2030 and 2100 were obtained by adding day and night +1°C to year 2030 and +4°C to year 2100 to the outside temperature of Haapastensyrjä breeding station in the year 2009. The temperature inside of the greenhouse never fell below of 0°C and the light hours of the material was not manipulated. Temperature sum of the Year 2100 had 2355 days degrees and for Year 2030 had 1615 dd, and the original temperature sum in the Year 2009 was 1270 dd. In the greenhouse the moisture was also controlled to obtained different growth condition and simulates the effects of this variable in the experiment. They were controlled by two different water treatments; the wet moisture condition with 700 mm/m 2 /year and the dry moisture condition with 260 mm/m 2 /year. The quantity of water per week was 7.0 l/box for wet treatment and 1.75 l/box for dry treatment. These values correspond to double amount of water for the wet moisture and 50% less to the dry moisture compared to the long term local average rainfall measured in Jokioinen (research station). The irrigation process was manually using a hosepipe in both treatments. Wet moisture material was irrigated twice per week and dry material was only once per week. The situations of boxes were changed randomly three times to minimize the location effects. For each moisture and temperature conditions, the same replications were done with the 9 different seeds origins for dormancy and 8 for dehardening experiment. For Material and Methods 21 each origin were done 4 and 5 replications, for dehardening and dormancy experiment respectively, according to the number of freezing tests necessary for the experiment. And finally, each freezing test had 4 seedlings (Table 2 and 3). Table 2. Number of oak material in dormancy experiment. Dormancy Temperature Moisture Origin Freezing tests Total seedlings Year 2100 Wet 9 5 180 Dry 9 5 180 Year 2030 Wet 9 5 180 Dry 9 5 180 Total 720 Table 3. Number of oak material in dehardening experiment. Dehardening Temperature Moisture Origin Freezing tests Total seedlings Year 2100 Wet 8 4 128 Dry 8 4 128 Year 2030 Wet 8 4 128 Dry 8 4 128 Total 512 For each condition used, i.e. for each combination of temperature, moisture, origin and finally the different freezing test dates, were sowed four seedlings replicates. 2.2.2 Growth conditions during dormancy experiment After the summer growth and hardening testing all the material was transferred to outside on the day 16.11.2009. The place was partly covered with plastic, but had open ceilings. The material in these conditions did not have any temperature control and could suffer a lot of minus temperatures. But the seedlings should acclimate correctly and be in deep dormancy. When the dormancy testing started, the boxes from the outside cold storages were taken into the same greenhouse of previous summer growth. In the greenhouses the temperature never was less than +2ºC. To evaluate the effect of frost in relation with Material and Methods 22 the length of high temperature treatment, the material was subject to one freeze testing per week. 2.2.3 Growth conditions during dehardening experiment In dehardening experiment all the material was in outside condition for the purpose of getting a natural timing of dehardening as the length of high temperatures and night length. Like in dormancy experiment, the material was subject to one freeze testing per week, corresponding the first material tested with a night length of 8hours and 25 minutes and the last one a night length of 6 hours and 25 minutes. 2.3. Freeze testing The freezing tests were made to simulate the frost occurring in the hibernation periods in order to ascertain the effects produced on the seedlings when they are in the state of dormancy and dehardening. The freezing test was an operation that affects the variables analyzed such as injuries, survival and height growth. All material used was submitted to the freezing test and were made different replications with the same conditions (seed origins, temperature and humidity) in order to ascertain the importance of the frost effect on seedlings depending on the day on which frost occurs. The freezing test experiment period started the last week of 2009 and ended the 8th week of 2010 for the dormancy experiment and for the dehardening experiment started the 17th week of 2010 and ended the 20th week of 2010. Freezing test process was consisting to transfer the material to be tested in a cold chamber with an automatically controlled temperature. The original temperature of the chamber was 5°C and from 21:00 the temperature cooled down constantly 3 º C per hour to reaching a minimum temperature of -10 º C at 2:00. During next two hours the temperature remained constant and then begins to warm up 3°C/h for 5 hours until 9:00 when it reached the initial temperature (Fig 5). After the freezing test the boxes were returned to their original condition in the greenhouse. Material and Methods 23 Figure 5. Freezing test process; temperature variation in the time. 2.4. Data processing First, the entries that had not germinated or were dead before the freeze test were removed. • Height before the freeze test: We measured the height of the seedlings before they were subjected to the freeze test. • The material was subjected to the freeze test according to the protocol described above. • The percentage of seedlings survival after the freezing test was calculated. • The before and after freezing test high data of seedlings that were used to calculate the percentage of seedlings injuries produced after the freezing test. 2.5. Statistical analysis Once the experimental data was obtained, the data was processed to determine the source effects; temperature, moisture, origin and freeze test day. Those sources have direct influence on the plants growth and development, the survival and the injuries in the dormancy and dehardening process. The statistical software used to perform every analysis was the SYSTAT 9 for Windows. With this program it was obtained all necessary statistical information to get the results, as well as to do the relevant tables, graphs and figures. The Analysis of Variance (ANOVA), was used to find out the results of the various models raised. In the Time 21:00 02:00 04:00 09:00 Tª variation cooling -3°C /h warming +3°C /h Temperature +5°C -10°C +5°C Material and Methods 24 analysis, on one hand, the independent variables were: temperature condition, moisture condition, seeds origin and time of freeze testing. On the other hand, the dependent ones were: the seedling height, percentage of survival, percentage of injuries and the time of bud formation. From this original data, there were summarized the seedling height average and as well as the survival and injuries percentage. It was reported that measures of variability using the standard deviation with the mean (mean ± SE). Also, all bar graph figures had a standard error of 0.6825 as an error bar. Correlations analysis between the height before freeze testing and injuries level were done using the type Pearson and corrected by the probabilities of Bonferroni. For best results it was used for the following data transformations: The parameters box edge and table edge representing the edge effect of the boxes and tables were created respectively, in the different variables to analyze. This effect is due to a competition between the seedlings depending on the location in the box, and also that irrigation could reach differently depending on the location of the box at the tables. This might affect the values of variables that are analyzed. The freezing test did not work properly in all sessions, so we had to remove some material with the objective of improve the analysis. Due to the lack of entries required to create the survival ANOVA model, a new variable that grouped the seedlings entries into two largest groups; Finland and South was created in both experiments. • Dormancy experiment: The time of freeze testing became in treat time because the first week of treatment was the number 51 and therefore, it would have been an error in the testing data freeze, so this was named treat time 0. In survival model, the origin North Poland was deleted because they had insufficient inputs to make a quality analysis. • Dehardening experiment: In some models, few origins were eliminated because they had insufficient inputs to make a quality analysis. - Injuries: The origin North Poland in dehardening experiment was deleted. - Survival model: The origin North Poland in dormancy experiment and North Poland and South Swedish in dehardening experiment were deleted. Results 31 1250-1Malmi 1250-2Annala 1250-3Katari 1250-4Ruissa 1250-5Tenho 1250-6Parain 1550-7Swede 1650-9Latvia Origin (days degrees) 0 20 40 60 80 100 120 Survival (%) Year 2100 Year 2030 Temperature condition Figure 10. The mean percentage of survival in oak seedlings according of the origin (day degrees) and clustered by growth condition (Year 2030 and Year 2100). The factors temperature condition and seedling origin was significant, but on the contrary any interactions were significant (Table 7). Table7. Effect of previous summer temperature and moisture conditions together with oak origins and the time of testing into the survival of seedlings during dormancy period. Source Sum - of - Squares df Mean - Square F - ratio P Origin 5289.248 7 755.607 3.032 0.008 Temperature 1222 8.261 1 12228.261 49.073 0.000 Freeze testing week 1092.205 4 273.051 1.096 0.366 O * T 2687.679 7 383.954 1.541 0.169 O * F 7659.023 28 273.537 1.098 0.368 T * F 1030.966 4 257.742 1.034 0.396 T * F * O 7120.331 28 254.298 1.021 0.457 Error 16944.4 44 68 249.183 Results 32 3.4. Dehardening 3.4.1. Injuries The seedlings injury augmented in all growth condition from 5.2% (± 4.4%) of average in the first freeze testing in the week 17 to the week 19 with a 59.1% (± 4.6%) of average. After this trend, the injury direction decreased in the last freeze testing week, the average was 52.6% (±5.0%), except the growth condition Year 2030 dry where injuries still increase (Figure 11). 16 17 18 19 20 21 Freeze test week 0 10 20 30 40 50 60 70 80 90 100 Seedling injury (%) Year2100 - wet Year2100 - dry Year2030 - wet Year2030 - dry Growth conditions Figure 11. The mean percentage of injuries in oak seedlings according of the freeze testing (weeks) and clustered by growth condition (Year 2030 and Year 2100). Temperature of treatment (p <0.05), the moisture condition (p<0.05) and their interaction (p<0.05) explained significantly injuries variation (Table 8) Seedlings grown in Year 2030 condition during previous summer had significantly lower injuries level than seedlings grown in more warm Year 2100 conditions. There were less injuries in dry moisture condition than in the wet condition. The minimum injuries were 13.4% (±4.7%) for Year 2030 dry and 35.2% (±4.3%) for the wet condition. Otherwise, the maximum injuries were 38.8% (± 4.9%) for the Year 2100 dry and 41.5 (±4.3%) for the wet condition (Figure 12). Results 33 Year2030 Year2100 Temperature condition 0 10 20 30 40 50 60 70 80 90 100 Seedling injury (%) Wet Dry Moisture condition Figure 12. The mean percentage of oak seedlings injuries according of the temperature condition (Year 2030 and Year 2100) and clustered by moisture condition (wet and dry). Variations of injuries were explained by previous summer temperature, freeze testing week and moisture (Table 8). The variable large origin was not significant (p > 0.05) as the differences were less than 0.5 %, thus there were not differences between the Finnish and more Southern seedlings. There were also some significant interactions to explain the variability; temperature* freeze testing week, temperature*moisture, freeze testing week*moisture, large origin*temperature* freeze testing week (Table 8). The effect of the position of the box in the table was significant (p<0.05), but not the seedling position in the boxes. Results 34 Table 8. Effect of previous summer temperature and moisture conditions together with oak origins and the time of testing into the injuries of seedlings during dehardening period. Source Sum - of - Squares df Mean - Square F - ratio P Large origin 4.065 1 4.065 0.004 0.947 Temperature 10783.164 1 10783.164 11.931 0.001 Freeze testing week 95580.102 3 95580.102 0.004 0.000 Moisture 6544.409 1 6544.409 7.241 0.008 O * T 3.657 1 3.657 0.004 0.949 O * F 503.597 3 167.866 0.186 0.906 O * M 891.104 1 891.104 0.986 0.321 T * F 22986.459 3 7662.153 8.478 0.000 T * M 4027.482 1 4027.482 4.456 0.036 F * M 8394.943 3 2798.314 3.096 0.027 O * T * F 9993.465 3 3331.155 3.686 0.012 O * T * M 938.541 1 938.541 1.038 0.309 O * F * M 4508.667 3 1502.889 1.663 0.175 T * F * M 6962.153 3 2320.718 2.568 0.054 O * T * F * M 2208.521 3 736.174 0.815 0.487 Box edge 1187.051 1 1187.051 1.313 0.253 Table edge 10015.377 1 10015.377 11.081 0.001 Error 283793.275 314 903.800 Results 35 3.4.2. Survival The average percentage of survival of the oak seedlings was 83.8% (± 2.7%) during dehardening experiment. Previous summer temperature condition did not explain (p>0.05) the survival variation (Table 9). The moisture was significant (p<0.05) with 88.2% (±3.8%) survival for the dry condition and 79.2% (± 3.8%) for the wet condition. In the first freeze testing week the survival was around 100% not depending much of previous summer growth condition. Then the survival decreased until the freeze testing week 19 with 70% (±4.2%) of survival. In the last freeze testing week the survival increased up to 79.5% (±4.2%) (Figure 13). 16 17 18 19 20 21 Freeze testing week 0 20 40 60 80 100 120 Survival (%) Year2100 - wet Year2100 - Dry Year2030 - wet Year2030 - Dry Growth condition Figure 13. The mean percentage of survival in oak seedlings according of the freeze testing (week) and clustered by growth condition (Year 2030 and Year 2100). The lowest survival seedling was the origin from Tenhola (Finland) with 68.8% (± 5.5%) of average and the highest average was from Ruissalo (Finland) with 95.8% (± 5.1%) of survival (Figure 14). Results 36 1250-1Malmi 1250-2Annala 1250-3Katari 1250-4Ruissa 1250-5Tenho 1250-6Parain 1650-9Latvia Origin (day degrees) 0 20 40 60 80 100 Survival (%) Year 2100 - wet Year 2100 dry Year 2030 - wet Year 2030 - dry Growth condition Figure 14. The mean percentage of survival of oak seedlings according of the origin (day degrees) and clustered by growth condition (Year 2030 and Year 2100). Survival variation was explained significantly by origin and freeze testing week and also their interaction was significant (Table 9). The temperature condition was not significant (P>0.05). Table9. Effect of previous summer temperature and moisture conditions together with oak origins and the time of testing into the survival of seedlings during dormancy period. Source Sum - of - Squares df Mean - Square F - ratio P Origin 9165.267 6 1527.544 3.134 0.010 Temperature 411.793 1 411.793 0.845 0.362 Freeze testing week 14906.174 3 4968.725 10.195 0.000 O * T 5709.005 6 951.501 1.952 0.089 O * F 1 8407.503 18 1022.639 2.098 0.018 T * F 1212.672 3 404.224 0.829 0.483 T * F * O 10135.403 18 563.078 1.155 0.329 Error 26805.556 55 487.374 Discussion 37 3.5. The effect of seedlings size on the injuries There were in dormancy experiment a negative correlation, r= -0.167 (p= 0.000, N= 849) between the seedlings size before the freeze testing and the results of the injuries. In origin level only Malmi, r= -0.203 (p= 0.041, N= 101) and Sweden, r = -0.326 (p= 0.001, N= 110) had significant correlations. There were also negative correlation r= -0.102 (p= 0.012, N=615) between the seedlings height and injuries in dehardening experiment. In origin level only Malmi, r= - 0.215 (p= 0.049, N=84) and Parainen, r= -0.208 (p= 0.001, N=110) had significant correlations. Discussion 38 4. DISCUSSION 4.1 Experimental process Greenhouses create problems because of difficulties in randomize tested material to obtain similar kind of growth conditions. Even though the tables were randomized three times during growing period the difficulties still persisted. One of the main disadvantages was due to the use of automated irrigation in greenhouses. This creates effect on the edge of the tables and boxes. This effect consists in unequal contribution of water by the irrigation system to the seedlings, depending on the location of the boxes in the tables as well as their own seedlings in each box. In the ANOVA model we could include these sources as covariates to determine their effects on the seedlings. The results of the table location, indicate that the error that provides the unequal amounts of water supplied by the irrigation system together with the competition of plants, influences in the height of the seedling. This also has an influence in the injuries level in dehardening experiment. This results show an error in the process of changing the boxes among the tables or in the methodology used. On contrary, in any case the boxes not were affected fort the water irrigation, so we can assume that seedlings position in the box have no effects. There is a relationship between the height of seedlings and the percentage of injuries that occur after the freeze testing, having a negative correlation. If the height of seedling decreased, the percentage of injuries inflicted by freeze testing tends to increase. The seedlings’ height depends on their origin. Thus, when analyzing the different parameters (injuries and survival) related to previous summer temperature, moisture and origin, they will be influenced by the height that also depends on their origins. Therefore, the origin is being evaluated twice when using this method. But this has little relevance since the correlation is not strong because it is close to 0. Experimental measurement of cold hardiness provides a useful means for examination of different hardiness characteristics in plants, and for defining the impact of environmental factors on the level of hardiness, being possible to test conditions not yet available in nature. For testing the cold hardiness level we used the freeze testing, being a good system to simulate the frost situation, but not the possible effects of the snow cover. Discussion 39 The fact of having used greenhouses to simulate the predicted different thermal, as well as the use of freeze testing, provides us a methodology easy to use and replicate. But this only limits to a laboratory level, and it can be different from the results obtained in field experimentation. All in all, Lindén (2002) points out that the results from laboratory tests can be used to supplement and accelerate field experimentation. 4.2 Effect of growth condition on oak seedling development 4.2.1 Survival after growth period The average of survival before freeze testing was about 74%, with differences of survival among the origins in excess of 25% units of the average. This is the case of Polish origins with only 48% survival seedlings (figure 6). Probably this was because they were collected in 2008, one year before than the rest of the acorns. García-Fayos (2001) asserts that acorns of Fagaceae family, from Mediterranean species, lose their viability quickly under ambient conditions, this fact could explain the low survival of Polish seedlings before the freeze testing. Excluding Poland origin, there were not significant differences among the origins from Finland and the other origins. 4.2.2 Height before freezing test The tallest seedlings were from the previous summer temperature condition of year 2030 with 13% higher than seedlings of year 2100 (Figure 7). In lower temperature summer, the oak growth taller (in south Finland) than in high temperature condition. This contrasts with Matala (2005) that show that in warming conditions produced by climate change the trees should growth taller than in lower summer temperature. It was seen that the seedlings from the wet growth condition was 8.2 % taller than dry condition. For the pedunculate oak seedlings, the lack of water in the dry conditions has a negative effect on their growth. Origins had a big influence to the height of the seedlings. The seeds from Finland, with a 1250 day degrees in nature growth conditions, had similar growth being the smallest. More Southern origins, was about 33% taller than Finland origins. The tallest origin was from Poland, which also has the highest day degrees, taking the double of height than Parainen. Significantly, there was a relationship between the growth of seedling and the day degree of the seed origin, taking taller seedlings in major temperature sum. Discussion 40 4.3 Dormancy state The survival of the seedlings to the winter depends on both the ability to reach the maximum capacity of cold hardening, and the timing and rate of cold acclimation, the level of cold tolerance during dormancy and the loss of hardening (Lennartsson 2003). When the oak seedlings is in deep dormancy then can tolerance temperatures about - 40ºC (Repo et al. 2008). In this experiment we took seedlings from deep dormancy and exposed them to warm conditions and noticed that they cannot tolerance even -10ºC temperatures without injuries. The previous summer growth temperatures seems also have an significant impact on the break-down of dormancy. With higher summer temperature, higher injuries during dormancy break down experiment. Moisture, by itself, had no significant effect. By the year 2030, the seedlings are in a complete state of cold hardiness, in which they are capable of bearing the simulated frosts, having null injuries and a survival of almost 100%. On the contrary, by year 2100 it can be observed that the frosts will have huge consequences in the deep dormancy, since there is a 50% rate of injuries and a 20% rate of mortality (Figure 8). Thus, comparing the results of both years, we conclude that the highest temperatures, during the period of growth, directly affect the risk of frost damage and survival during the dormancy state. There are variations in the survival among the different origins, but not in injuries, that as it is observed in the figure 10, largely owe to the low percentage of survival of Malmi and the high ones of Latvia, with a difference of 22% units. Nevertheless, the origin of Malmi is a park type, so it is possible that this fact could slant the results and finally the origin is not significant due to its weak value. References 47 Repo,T., Mononen, K., Alvila, L., Pakkanen, T. and Hänninen, H. 2008. Cold acclimation of pedunculate oak (Quercus robur L.) at its northernmost distribution range. Environmental and Experimental Botany 63: 59–70. SYSTAT Version 9. 1998. SPSS Software Inc. for Windows. Chicago. Talkkari, A. 1996. Regional predictions concerning the effects of climate change on forests in southern Finland. Silva Fennica 30(2-3): 247-257. UNFCC. 1999. The United Nations Framework Convention on Climate Change ( Article 1.2.). Vakkari, P., Blom, A., Rusanen, M., Raisio, J. and Toivonen, H. 2006. Genetic variability of fragmented stands of pedunculate oak (Quercus robur) in Finland. Genetica 127:231–241. White, J. and Walters, S. M. 2005. Trees: A field guide to the trees of Britain and Northern Europe.Oxford University Press. Xin, Z. and Browse, J. 2000. Cold comfort farm: the acclimation of plants to freesing temperatures. Plant, Cell and Environment 23: 893-902. Zanetto, A., Roussel, G. and Kremer, A. 1994. Geographic variation on inter-specific differentiation between Quercos robur l. and Quercus petraea (Matt.) Liebl. Forest Genetics 1(2):111-123.