Skip to main content

Temporal instability of tree-ring/climate relationships: tree responses to climatic change and implications for paleoclimate research

Final Report Summary - TREE-RINGS & CLIMATE (Temporal instability of tree-ring/climate relationships: tree responses to climatic change and implications for paleoclimate research.)

The proposed project number 253277 with acronym ‘Tree rings and climate’ aimed to provide a greater level of understanding of the temporal instabilities in the relationships between tree rings and climate documented around the world. This is particularly critical issue with potential implications for the global carbon cycle, forest growth patterns and climate reconstructions. This project has been addressing the impacts of climatic change on forests in terms of ecology, and also questions regarding climatology and atmospheric sciences as well. This project employed classical dendrochronological techniques and statistics in combination with an effort toward improving tree ring models and synthetic growth model predictions. The ecological component was based on the study of forest responses to changing climatic conditions, while the climatological component was assessing the strength of climatic signals in tree-ring data, with the goal being to reduce uncertainties in climatic reconstructions.
The study of the “divergence problem” detected on the boreal forests (aim 1.1 and aim 2) and the validation and/or improvement of temperature reconstructions for Alaska (aim 3.1) were scheduled and were conducted during the first 2 years corresponding to the outgoing phase of the project (USA). The stable isotopes analyses proposed for the aim 2 were replaced by densitometric analyses to measure the density of several portions of the tree rings. The switch from stable isotopes to density measurements was extremely fruitful for the achievement of the main goal of the proposal that seeks how to deal with temporal instabilities in the relationships between tree rings and climate. However, during the 3rd year, we decided to push forward this research using new parameters and using the remaining funds of the grant to perform isotopic measurements, δ13C and δ18O ratios, in wood available from the Firth River, which was the location studied in Alaska that lead to publications during the outgoing phase. The addition of the isotopes have provided some important data that clarified the physiological mechanism underneath the ‘divergence problem’ as was suggested in the initial proposal. The Aim 1.2 and Aim 3, related with the Iberian Peninsula, were accomplished during the 3rd year of the project. Some of this data is already published, whereas some other is currently in progress toward becoming additional publications. Overall, we have fully accomplished the aims proposed in the ‘Tree-Rings and Climate’ project.

In relation to Aim 1.1 main results were the assessment of the stability of the temperature climatic signal and the results were published in Environmental Research Letters, Andreu-Hayles et al. 2011. The focus was on the high frequency domain, thus only the residual TRW and MXD chronologies were chosen for comparison with the meteorological data. Since almost identical results were obtained when using different residual chronologies. The residual TRW chronology shows significant Pearson’s correlation coefficients for June and July Dawson mean temperatures for the 20th century. The correlation with summer temperature was highly significant for the first half of the 20th century, but non significant for the second half. Significant Pearson’s correlation were found between the residual MXD chronology and May, July and August Dawson mean temperatures for the 1901-2001 period. July-August temperatures were significant through the whole 20th century. Non-clear patterns were found when using precipitation or PDSI gridded products, whereas significant results were obtained when using gridded mean temperatures. Field correlations obtained using the CRU TS 2.1 or GISTEMP grid product were very similar. Field correlations between the TRW chronology and the MXD chronology versus May to August gridded mean temperatures confirm the results obtained with Dawson local station data for a larger spatial scale. A complete absence of temperature signal during the second half of the 20th century was assessed for TRW, whereas the July-August temperature signal remained for MXD during both studied periods. The fellow also assessed variations in growth and density. To assess the tree growth at the Firth River site, the raw tree-ring data was used after removing from all the series the first 150 years, as well as all the rings older than 300 years. Hence, only the information of trees when were holding this age range was considered. The TRW frequencies for different time periods were compared. A higher frequency of wider rings was observed from 1900 to 2002 in comparison with those prior to 1900 since 1400, as well as during the second half of the 20th century in comparison with the first half. The empirical CDF and the KS test of both comparisons demonstrated that TRW distributions were significantly different, showing a significant shift toward wider rings after 1900, and even wider after 1950. The MXD frequencies show rings with higher MXD values from 1900 to 2002 than prior to 1900 since 1400. Regarding the 20th century, MXD values were even higher during second half than during the first half. The empirical CDF and the KS test assessed that that the MXD distribution show significantly denser rings during the 20th century in comparison with previous centuries, and even denser after 1950.

In relation to Aim 2, the fellow generated tree-ring width and the Maximum Latewood density chronologies. This work was also reported in the Environmental Research Letters paper (Andreu-Hayles et al. 2011). Living and subfossil wood samples of white spruce were collected from the Firth River area of the Alaskan National Wildlife Refuge (ANWR) in 2002 and were available in the Tree Ring Lab when the fellow started her postdoctoral stay. These samples have yielded one of the very few millennial length tree-ring records presently available for northern Alaska. At this latitudinal treeline setting (68.78N 142.35W) trees are considered to have been strongly limited by temperatures over the past thousand years, at least until the recent anthropogenic period of recent decades. A tree ring-width chronology was established to identify the nature of forest response to climatic and environmental changes for white spruce (Picea glauca), a dominant Arctic treeline species. Windendro 2012 (Regent Instrumental Canada Inc.), software for tree-ring analyses, was used to generate new data and improve methodological procedures during this project. The Firth River tree-ring width chronology is based on 233 samples from 111 trees (31 living trees and 84 cross-sections from subfossil wood). It is a millennial-length record that spans from AD 1067 to 2002 (936 years). The raw ring-width measurements for the Firth site were standardized using the ARSTAN software for removal of nonclimatic trends. Standardization was performed on the combined living and subfossil wood samples using a negative exponential curve (NEXP) or Regional Curve Standardization (RCS) method, which optimizes retention of centennial-scale variability. Variance was stabilized using a power transform technique (based on local mean and standard deviation) that reduces potential end-fitting bias. The Maximum Latewood density (MXD) chronology, spanning from 1073 to 2002 AD, is based on 246 measurement time series from 30 living and 74 subfossil trees that were also used for building the TRW chronology.

Additionally, also related to Aim 2, stable isotopes were analyzed from tree-ring from Alaska. A total of 333 tree-ring wood samples from 3 trees from the Firth River in Alaska (Andreu-Hayles et al. 2011) were processed for measuring stable isotopes. δ13C and δ18O tree-ring chronologies spanning from 1900 to 2010 were generated and the dominant climatic factors controlling the isotopic ratios was found to be temperature. While δ18O ratios are highly affected by May-July temperatures, this signal is weaker in δ13C ratios. This suggests distinct environmental controls recorded by the two isotopes and that δ13C may be potentially more sensitive to other factors (e.g. photon flux) affecting photosynthetic rate than just temperatures. From observations alone, however, we can only use the correlation with different observational estimates to speculate on the relative importance of each, which is complicated by the fact that they will often co-vary. Formal attribution requires mechanistic models with the ability to estimate, through sensitivity experiments, the complex physiological response to environmental changes.

In relation to Aim 3.1 main result was a summer temperature reconstruction for the last nine centuries (Anchukaitis et al. 2012). The fact that the MXD chronology was tracking very well July and August mean temperatures throughout the 20th century, lead to the reconstruction of summer temperatures from Northwestern North America for the last nine centuries. The warmest period occurred during the late 20th century, most likely due to the rapid increases in greenhouse gases. Prior to this epoch during the preindustrial time (1100-1850 CE), temperatures fluctuations were in agreement with natural radiative forcing. It is widely known that temporal instabilities between tree ring proxies and climate are introducing a higher level of uncertainty in paleoclimatic reconstructions.

Finally, main results related to Aim 1.2 and Aim 3.2 were based on the identification of distinct regional and seasonal patterns of atmospheric variability using tree rings and stable isotopes from the Iberian Peninsula and links to historical archives. This study seeks to validate different approaches to understand the relative contribution of Mediterranean and Atlantic atmospheric influences on several sites across the Iberian Peninsula for the last 400 years combining the information recorded by both tree rings and historical archives with field correlation and climate dynamical analyses. Measuring stable carbon (δ13C) and oxygen (δ18O) isotope ratios, which is more expensive and time-consuming than measuring the classical ring-width parameter, provides an added value by better capturing large-scale climatic features than ring-width, which is often more dependent on local conditions. Spatial correlations between tree rings and gridded climate products demonstrate that the isotope signatures in the targeted Iberian pine forests are very sensitive to water availability during the summer period because they are strongly dominated by stomatal conductance. In agreement, composite anomalies during years with extreme high (low) stable isotopic values in the tree-ring records displayed coherent large-scale atmospheric circulation patterns with reduced (enhanced) moisture flux for the different sites. Moreover, these analyses of extremes revealed that high/low proxy values do not necessarily correspond to mirror images in the atmospheric anomaly patterns, suggesting different drivers of these patterns and the corresponding signature recorded in the proxies. Therefore, this non-linear approach provides information not available with standard techniques. This approach was used in multiple comparisons among tree rings, instrumental/reanalysis products, and historical archives to extend our analyses for different seasons to past centuries.

In conclusion, the ‘Tree rings and climate’ project reported evidences to support the statement that applying dendrochronological methods can provide a long-term context back in time from several centuries to millennia with an annual resolution only available when using paleorecord such as tree rings. At present, dendroclimatology, based on the assumption that it is possible to extract climatic information registered in tree-rings formed in the past, face significant challenges in understanding and attributing the causes of phenomena such as the “divergence problem”, and in finding the correct way to estimate past climate successfully dealing with these temporal instabilities. This project successfully addressed crucial questions in relation to climatic change research by mean of analyzing different tree-ring parameters and applying different statistical approaches, overcoming the limitation previously described.

Socio-economic impact of the project
At the European level, climate change currently constituted one of the main areas of research during the cooperation programme of FP7 because it was considered that “a multidisciplinary and integrated research is required in order to advance our knowledge on the interactions between the climate, biosphere, ecosystems and human activities for a sustainable management of the environment and its resources”. The multidisciplinary project developed here targeted perfectly this aim, studying the interactions between climate and forests to assess and predict forest responses to the changing climatic conditions, but also to extract the climatic signal registered in tree rings. This project provided expertise and know-how in this research area in two important ways: tree-growth and proceeding to reduce uncertainties in paleoclimate reconstructions. The number of research groups in Europe specifically dedicate to tree-ring research is small. If we further restrict our count to the groups that have the capacity of generating tree-growth models, stable isotopes and other new procedures for the improvement of dendrochronological techniques classically used in climate research, the number drops even further. Therefore, the multidisciplinary study on temporal instability on trees-climate relationship that the fellow conducted during her on-going project constituted a significant step forward in the European leadership in the field. Importantly, the acquired know-how was transferred to the newly created Catalan Institute of Climate Research (IC3), where the returning phase of the fellowship was conducted.
The uncertainty in paleoclimate reconstructions is considered the main obstacle for a good comprehension of past climate variability that is a key factor for predictions of future climate change. On the other hand, ecosystems and species are very likely to show a wide range of vulnerabilities to climate change, pinpointing the boreal forests and the Mediterranean-climate ecosystems as one of the most vulnerable with a high confidence level. The accomplishing of this project has been highly beneficial for European communities living in the Mediterranean region providing information that will help to mitigate the ongoing climate changes expected to intensify drought stress in this region affecting water resources, and so food security, biodiversity, health, economy, amongst other factors. Climate change and its impacts are one of the greatest environmental, social and economic threats that our planet is facing, fast becoming a key international problem of the 21st century. Therefore, the developed activity will help to quantify local impacts of climate change in the most sensitive regions of Europe (Iberian Peninsula) and worldwide (Boreal region), underpinning further policy options. Research on climate change processes and impacts on natural resources and humankind will help to identify and assess key drivers and improve understanding of their interactions.