NAm2kDendro

By McPartland et al. | May 22, 2024

Introduction

The NAm2kDendro database contains 230 tree ring records (both tree ring width and maximum latewood density) from North America. Each chronology has been detrended using six distinct procedures to facilitate sensitivity analyses. The database also includes additional metadata, such as the Expressed Population Signal (EPS) and the number of individual tree cores contributing to the chronology at each timestep. This is a dynamic compilation, meaning that new datasets that meet the criteria for inclusion can be added and included in subsequent versions. See the criteria and data submission sections below for details.

Data

Data access and LiPDverse visualizations are available here.

Publication

The paper describing this compilation was published in Geophysical Research Letters and is available here.

How to cite this compilation

McPartland, M. Y., Dolman, A. M., & Laepple, T. (2024). Separating common signal from proxy noise in tree rings. Geophysical Research Letters, 51, e2024GL109282. https://doi.org/10.1029/2024GL109282

Detrending Techniques

Tree-ring chronologies are standardized (“detrended”) to remove non-climatic noise, primarily the biological growth trend related to the age and size of the tree. The NAm2kDendro database provides versions of each chronology processed with six different detrending techniques to allow users to assess the sensitivity of climate reconstructions to these methodological choices. In the database, the “detrendingMethod” field at the column level describes which method was used for each chronology.

  • Negative Exponential (NegEx) : This is a classic detrending method that fits a decreasing exponential curve to each tree-ring series. This curve represents the idealized biological growth trend, which is then removed from the raw measurements, leaving behind a series of growth indices that are presumed to reflect environmental influences.

  • Regional Curve Standardization (RCS): This method aligns all individual tree-ring series from a site by their biological (cambial) age and averages them to create a single “regional curve”. This curve, which represents the common age-related growth trend for the site, is then used as the detrending function for each individual series. RCS is particularly effective at preserving low-frequency climate signals when sample replication is high throughout the chronology.

  • Age Dependent Standardized Chronology (AgeDependentStdCrn): This method is a variation of Regional Curve Standardization (RCS) that explicitly models how the variance of tree-ring series changes with cambial age. By standardizing based on an age-dependent mean and variance, it aims to produce a more robust estimate of the common climate signal, especially in cases where sample replication is low.

  • Age Dependent Standardized Chronology with variance stabilization (AgeDependentStdCrnStb): This technique combines the ADSC method with an additional step to stabilize the variance of the final chronology over time. This addresses the common issue where the variance of a chronology decreases in periods with fewer contributing tree-ring series, ensuring that the variance is more uniform throughout the length of the record.

  • Signal-Free Detrending (SsfCr) : This is an iterative method designed to better preserve low-frequency (long-term) climate signals that might be inadvertently removed by other techniques. In the first pass, a master chronology (the “signal”) is created; in the second pass, this signal is temporarily removed from each individual series before detrending, and then added back to the resulting indices. This process helps to distinguish the common climate signal from the individual biological growth trend of each tree.

  • Signal-Free Detrending with variance stabilization (SsfCrnStb): This approach applies the same variance stabilization process described above to chronologies created using the Signal-Free Detrending (SFD) method. The goal is to produce a chronology that both preserves low-frequency climate signals and has a stable variance over time.

Usage notes

All 230 datasets that were restandardized for this compilation are included in the compilation, however not all are considered to be temperature-sensitive or appropriate for climate reconstructions.

  1. All included data are tree-ring chronologies (width and/or density) from North America that are at least 100 years long.
  2. Most of the datasets include raw ring width or density values, but unless you’re sure what you’re doing, you’ll want to use the chronologies for climate analysis or reconstruction. The chronology variables are trsgi (tree ring standardize growth index) or trmxdsgi (tree ring maximum latewood density standardized growth index). Raw values are noted as ringWidth and MXD.
  3. Every chronology was evaluated for temperature sensitivity by correlating with local temperature from CRUTEM 4.3.0.0. Only those with siginicant positive correlations are labeled with temperature interpretations
  4. The Expressed Population Signal (EPS) was also screened for all chronologies, and only those that pass this screening and that are marked as temperature sensitive are marked as primary timeseries.
  5. Many datasets have multiple chronologies that pass these screens and so have multiple chronologies that are marked as primary. Consider carefully which detrending approach makes the most sense for your application.
  6. This dataset likely has uses beyond temperature reconstruction, and some non-temperature climate-sensitive chronologies are included.

Long-term archival

The NAm2kDendro data are archived at the NOAA National Centers for Environmental Information (NCEI) World Data Service for Paleoclimatology, and are available here.

Funding

Funding to compile and curate NAm2KDendro compilation was originally provided by the USGS Powell Center for Synthesis and Analysis, and this work was conducted by Drs. Kevin Anchukaitis, Nick McKay, Greg Pederson, Cody Routson, Scott St. George and Uday Thapa