Search

Forecasts of future forest change are governed by ecosystem sensitivity to climate change, but ecosystem model projections are under‐constrained by data at multidecadal and longer timescales. Here, we quantify ecosystem sensitivity to centennial‐scale hydroclimate variability, by comparing dendroclimatic and pollen‐inferred reconstructions of drought, forest composition and biomass for the last millennium with five ecosystem model simulations. In both observations and models, spatial patterns in ecosystem responses to hydroclimate variability are strongly governed by ecosystem sensitivity rather than climate exposure. Ecosystem sensitivity was higher in models than observations and highest in simpler models. Model‐data comparisons suggest that interactions among biodiversity, demography and ecophysiology processes dampen the sensitivity of forest composition and biomass to climate variability and change. Integrating ecosystem models with observations from timescales extending beyond the instrumental record can better understand and forecast the mechanisms regulating forest sensitivity to climate variability in a complex and changing world. [ABSTRACT FROM AUTHOR]

Reconstructions of prehistoric vegetation composition help establish natural baselines, variability, and trajectories of forest dynamics before and during the emergence of intensive anthropogenic land use. Pollen–vegetation models (PVMs) enable such reconstructions from fossil pollen assemblages using process-based representations of taxon-specific pollen production and dispersal. However, several PVMs and variants now exist, and the sensitivity of vegetation inferences to PVM selection, variant, and calibration domain is poorly understood. Here, we compare the reconstructions, parameter estimates, and structure of a Bayesian hierarchical PVM, STEPPS, both to observations and to REVEALS, a widely used PVM, for the pre–Euro-American settlement-era vegetation in the northeastern United States (NEUS). We also compare NEUS-based STEPPS parameter estimates to those for the upper midwestern United States (UMW). Both PVMs predict the observed macroscale patterns of vegetation composition in the NEUS; however, reconstructions of minor taxa are less accurate and predictions for some taxa differ between PVMs. These differences can be attributed to intermodel differences in structure and parameter estimates. Estimates of pollen productivity from STEPPS broadly agree with estimates produced for use in REVEALS, while comparison between pollen dispersal parameter estimates shows no significant relationship. STEPPS parameter estimates are similar between the UMW and NEUS, suggesting that STEPPS parameter estimates are transferable between floristically similar regions and scales. [ABSTRACT FROM AUTHOR]

High throughput phenotyping technologies are lagging behind modern marker technology impairing the use of secondary traits to increase genetic gains in plant breeding. We aimed to assess whether the combined use of hyperspectral data with modern marker technology could be used to improve across location pre-harvest yield predictions using different statistical models. A maize bi-parental doubled haploid (DH) population derived from F1, which consisted of 97 lines was evaluated in testcross combination under heat stress as well as combined heat and drought stress during the 2014 and 2016 summer season in Ciudad Obregon, Sonora, Mexico (27°20" N, 109°54" W, 38 m asl). Full hyperspectral data, indicative of crop physiological processes at the canopy level, was repeatedly measured throughout the grain filling period and related to grain yield. Partial least squares regression (PLSR), random forest (RF), ridge regression (RR) and Bayesian ridge regression (BayesB) were used to assess prediction accuracies on grain yield within (two-fold cross-validation) and across environments (leave-one-environment-out-cross-validation) using molecular markers (M), hyperspectral data (H) and the combination of both (HM). Highest prediction accuracy for grain yield averaged across within and across location predictions (rGP) were obtained for BayesB followed by RR, RF and PLSR. The combined use of hyperspectral and molecular marker data as input factor on average had higher predictions for grain yield than hyperspectral data or molecular marker data alone. The highest prediction accuracy for grain yield across environments was measured for BayesB when molecular marker data and hyperspectral data were used as input factors, while the highest within environment prediction was obtained when BayesB was used in combination with hyperspectral data. It is discussed how the combined use of hyperspectral data with molecular marker technology could be used to introduce physiological genomic estimated breeding values (PGEBV) as a pre-harvest decision support tool to select genetically superior lines. [ABSTRACT FROM AUTHOR]

The construction of accurate age–depth relationships and a realistic assessment of their uncertainties is one of the fundamental prerequisites for comparing and correlating late Quaternary stratigraphical proxy records. Four widely used age–depth modelling routines – CLAM, OxCal, Bacon and Bchron – were tested using radiocarbon dates simulated from varved sediment stratigraphies. All methods produce mean age–depth models that are close to the true varve age, but the uncertainty estimation differs considerably among models. Age uncertainties are usually underestimated by CLAM, whereas age uncertainties produced by Bchron are often too large. With OxCal and Bacon, the setting of model-specific parameters influences the estimated uncertainties, which vary from too large to too small. The variability of sediment accumulation rates is underestimated by CLAM but overestimated by Bacon and Bchron. Bayesian age–depth models mainly improve the assessment of uncertainties of age–depth models. [ABSTRACT FROM AUTHOR]

Reconstructions of summer, winter or annual mean temperatures based on the species composition of bioindicators such as pollen, foraminifera or chironomids are routinely used in climate model-proxy data comparison studies. Most reconstruction algorithms exploit the joint distribution of modern spatial climate and species distribution for the development of the reconstructions. They rely on the space-for-time substitution and the specific assumption that environmental variables other than those reconstructed are not important or that their relationship with the reconstructed variable(s) should be the same in the past as in the modern spatial calibration dataset. Here we test the implications of this "correlative uniformitarianism" assumption on climate reconstructions in an ideal model world, in which climate and vegetation are known at all times. The alternate reality is a climate simulation of the last 6000 years with dynamic vegetation. Transient changes of plant functional types are considered as surrogate pollen counts and allow us to establish, apply and evaluate transfer functions in the modeled world. We find that in our model experiments the transfer function cross validation r2 is of limited use to identify reconstructible climate variables, as it only relies on the modern spatial climate-vegetation relationship. However, ordination approaches that assess the amount of fossil vegetation variance explained by the reconstructions are promising. We furthermore show that correlations between climate variables in the modern climate-vegetation relationship are systematically extended into the reconstructions. Summer temperatures, the most prominent driving variable for modeled vegetation change in the Northern Hemisphere, are accurately reconstructed. However, the amplitude of the model winter and mean annual temperature cooling between the mid- Holocene and present day is overestimated and similar to the summer trend in magnitude. This effect occurs because temporal changes of a dominant climate variable, such as summer temperatures in the model's Arctic, are imprinted on a less important variable, leading to reconstructions biased towards the dominant variable's trends. Our results, although based on a model vegetation that is inevitably simpler than reality, indicate that reconstructions of multiple climate variables based on modern spatial bio-indicator datasets should be treated with caution. Expert knowledge on the ecophysiological drivers of the proxies, as well as statistical methods that go beyond the cross validation on modern calibration datasets, are crucial to avoid misinterpretation. [ABSTRACT FROM AUTHOR]

We examine the ability of four different regression-tree ensemble techniques (bagging, random forest, rotation forest and boosted tree) in calibration of aquatic microfossil proxies. The methods are tested with six chironomid and diatom datasets, using a variety of cross-validation schemes. We find random forest, rotation forest and the boosted tree to have a similar performance, while bagging performs less well and in several cases has trouble producing continuous predictions. In comparison with commonly used parametric transfer-function approaches (PLS, WA, WA-PLS), we find that in some cases tree-ensemble methods outperform the best-performing transfer-function technique, especially with large datasets characterized by complex taxon responses and abundant noise. However, parametric transfer functions remain competitive with datasets characterized by low number of samples or linear taxon responses. We present an implementation of the rotation forest algorithm in R. [ABSTRACT FROM AUTHOR]

Conventional cross validation schemes for assessing transfer-function performance assume that observations are independent. In spatially structured environments this assumption is violated, resulting in over-optimistic estimates of transfer-function performance. H-block cross validation, where all samples within h kilometres of the test samples are omitted, is a method for obtaining unbiased transfer-function performance estimates. In this study, we assess three methods for determining the optimal h. Using simulated data, we find that all three methods result in comparable values of h. Applying the three methods to published transfer functions, we find they yield similar values for h. Some transfer functions perform notably worse when h-block cross validation is used. [ABSTRACT FROM AUTHOR]

Reconstructions of summer, winter or annual mean temperatures based on the species composition of bio-indicators such as pollen, foraminifera or chironomids are routinely used in climate model-proxy data comparison studies. Most reconstruction algorithms exploit the joint distribution of modern spatial climate and species distribution for the development of the reconstructions. They rely on the assumption of "uniformitarianism", which implies that environmental variables other than those reconstructed should not be important, or that their relationship with the reconstructed variable(s) should be the same in the past as in the modern spatial calibration dataset. Here we test the implications of uniformitarianism on such reconstructions in an ideal model world, in which climate and vegetation are known at all times. The alternate reality is a climate simulation of last 6000 years with dynamic vegetation. Transient changes of plant functional types are considered as surrogate pollen counts, and allow to establish, apply and evaluate transfer functions in the modeled world. We find that the transfer function cross-validation r2 is of limited use to identify reconstructible climate variables, as it only relies on the modern spatial climate/vegetation relationship. However, ordination approaches that assess the amount of fossil vegetation variance explained by the reconstructions are promising. We furthermore show that correlations between climate variables in the modern climate/vegetation relationship are systematically extended into the reconstructions. Summer temperatures, the most prominent driving variable for modelled vegetation change in the Northern Hemisphere, are accurately reconstructed. However, the amplitude of the winter and mean annual temperature cooling between the mid-Holocene and present day is overestimated, and similar to the summer trend in magnitude. This effect occurs, because temporal changes of a dominant climate variable, such as summer temperature, are imprinted on a less important variable, leading to reconstructions biased towards the dominant variable's trends. Our results indicate that reconstructions of multiple climate variables from the same bio-indicator dataset should be treated with caution. Expert knowledge on the eco-physiological drivers of the proxies, and statistical methods that go beyond the cross-validation on modern calibration datasets are crucial to avoid misinterpretation. [ABSTRACT FROM AUTHOR]

Late Glacial and Holocene environmental changes were reconstructed using physical, chemical and biological proxies in Lake Myklevatnet, Allmenningen, (5º13′17″E, 61º55′13″N) located at the northern side of Nordfjorden at the coast of western Norway. Myklevatnet (123 m a.s.l.) lies above the Late Glacial marine limit and contains sediments back to approximately 14,300 years before a.d. 2000 (b2k). Because the lake is located ~48 km beyond the margin of the Younger Dryas (YD) fjord and valley glaciers further inland, and did not receive glacier meltwater from local glaciers during the YD, the lake record provides supplementary information to Lake Kråkenes that received glacial meltwater from a local YD glacier. Lake Myklevatnet has a small catchment and is sensitive to Late Glacial and Holocene climate and environmental changes in the coastal region of western Norway. The age-depth relationship was inferred from a radiocarbon- and tephra-based smoothing-spline model with correlated ages from oxygen isotope maxima and minima in the Late Glacial sequence of the NGRIP ice core (in years b2k) to refine the basal chronology in the Myklevatnet record. The results indicate a two-step YD warming, colder early YD temperatures than in the later part of the YD, and considerably more climate and environmental variability during the late Holocene in western Norway than recorded previously in the oxygen isotopes from Greenland ice cores. The Myklevatnet record is also compared with other Late Glacial and Holocene terrestrial and marine proxy reconstructions in the North Atlantic realm. [ABSTRACT FROM AUTHOR]

Reflectance spectroscopy in the visible spectrum (VIS-RS) is a method that has been successfully applied for inferring organic content of sediments. In this study, we test the applicability of VIS-RS to lake sediments in Norway. On the one hand we use conventional, established algorithms for inferring organic content of sediments, on the other hand we test the potential of multivariate calibration techniques to infer organic content. For absolute quantification of organic content, conventional C org measurements are needed when using conventional algorithms as well as when employing multivariate calibration techniques. Both, conventional algorithms and multivariate calibrations, result in estimates of organic content closely mirroring loss-on-ignition measurements. When using multivariate calibration techniques, a conventional C org measurement every 5 cm is sufficient to obtain estimates of organic matter that are more accurate than those obtained by means of conventional algorithms. Therefore, the potential of multivariate calibration techniques and VIS-RS to substitute measurements of more time consuming and costly sediment parameters (e.g. clay minerals) should be tested. [ABSTRACT FROM AUTHOR]

Chironomids were used to reconstruct mean July air temperatures between c. AD 1580 and 2001 at Lake Silvaplana, a varved lake located in the Engadine, eastern SwissAlps. The goal of this study was to reconstruct temperature changes at near-annual resolution, and validate the reconstruction by comparison with records based on early instrumental data, documentary proxy evidence, dendrochronology, geochemical (biogenic silica (BSi)) and mineralogical data (quartz/mica ratios) at local and regional scales. Warmer than-the-climatenormal (AD 1961-1990) mean July air temperatures were inferred between c. AD 1610 and 1662, AD 1710 and 1740, AD 1790 and 1866, AD 1940 and 1960 and AD 1990 and 2001. Colder-than-the-climate-normal July air temperatures were reconstructed between c. AD 1662 and 1710, AD 1740 and 1790, AD 1866 and 1919, and AD 1970 and 1990. The 420-year chironomid-inferred mean July air temperature record was significantly (p < 0.01) related to June-September (JJAS) temperatures reconstructed from early instrumental and documentary data at regional scale, JJA temperature inferred from documentary proxy evidence at local scale and summer temperatures based on early instrumental data in central Europe. When the Z-scores of warm/cold periods were compared between records, only one period (c. AD 1740-1790) did not show significant correlations between the chironomid record and any of the eight other records considered here, probably because of increased precipitation and changes in the sediment composition which influenced the chironomid assemblages. 75% of the periods considered had significant correlations between the chironomid records, and both the reconstruction based on quartz/mica ratios and the inferred JJAS early instrumental and documentary proxy evidence, while 60% of the periods showed significant correlations between the chironomid-based record and the reconstruction based on early instrumental data of Central Europe. These results suggest that chironomids in the sediment of Lake Silvaplana yield valid temperature reconstructions at regional scales for the last 420 years. [ABSTRACT FROM AUTHOR]