Next page



Charster is designed to ease the exploratory analysis of lake-sediment or bog/peatland charcoal records when the purpose is to identify peaks in the charcoal record that result from nearby fires. Charster provides a few statistical tools that help characterize the variability in the charcoal record and whether identification of peaks is warranted. Charcoal time series may be manipulated in various ways before placing a threshold to identify peaks.

Charster is most helpful for analysis of charcoal records that display distinct peaks. Recent collaboration among several labs has led to agreement on the conditions that yield such a record (see also Whitlock and Larsen 2002; Whitlock and Anderson 2003):

  1. Sampled contiguously, typically at least every 1 cm, corresponding to deposition times of 30 years or less.
  2. The fire regime is of infrequent but intense stand-replacing fires. Such a regime will deliver charcoal to lakes in infrequent pulses such that subsequent delivery to the lake bottom and mixing does not remove the fire signal.
  3. Lakes or bogs are small in size (< ca. 10 ha), and if lake sediment is analyzed, that the lakes are fairly deep (>2 m) to limit sediment disturbance.
  4. Charster requires age estimates on the top and base of each sample where charcoal was quantified. It is not possible to use the software on undated records.

The general approach of this analysis is to "decompose" a time series into two components: a slowly varying background and a high frequency peak component (Long et al., 1999). One may make different assumptions regarding the behavior of these two components (i.e., nature of variability over time, relationship to local fires, etc.), prompting different methods of analysis. The purpose of Charster is to allow one to easily explore these options and their impact on the final interpretation.

It is recommended that first-time users read through these help pages before delving into analyzing a record. Please email me with any suggestions.

Some analyses are described in:  Gavin, D.G., F.S. Hu, K.P. Lertzman and P. Corbett. 2006. Weak climatic control of forest fire history during the late Holocene. Ecology 87:1722-1732.

Comparison with CharAnalysis

CharAnalysis, by Phil Higuera, grew out of the analyses in Charster.  CharAnalysis provides many more diagnostic statistics for evaluating peaks in charcoal records, including the ability to define thresholds using local criteria and analyze fire-interval distributions.  Some other practical differences between the programs are 1) Data input format and treatment of missing data-gaps, 2) Charster's capacity to do simple analyses on non-interpolated data (while interpolation is necessary for more complex analyses in Charster and is the default in CharAnalysis), and 3) Many useful graphs produced by CharAnalysis.


This software was written by Daniel Gavin. Please read my software disclaimer.


Charster is based on CHAPS (Charcoal Analysis Programs) by Patrick Bartlein.

The United States Geological Survey (USGS) funded this work. Craig Allen greatly facilitated this project. Critical feedback was provided by Jenn Marlon and Phil Higuera.

The software is written in REALBasic, and the REALBasic project is available for those who wish to inspect the code.

updated 25 April 2008