Skip to content

Latest commit

 

History

History
23 lines (19 loc) · 2.19 KB

NEWS.md

File metadata and controls

23 lines (19 loc) · 2.19 KB
Raw

ShellChron 0.3.5

0.2.8

Added escape statements (on.exit()) to functions that change the WD Updated description to refer to published manuscript with more details on the methodology.

0.2.7

Removed characters that caused problems with building PDF manual Minor spelling checks Updated some formulae to make checking run more smoothly

0.3.5

Following peer-review, I have implemented options in ShellChron that enable the user to set the parameters of the SCEUA algorithm and disable the sinusoidal regression (sinreg). Previous versions of ShellChron underestimated the seasonal d18Oc range and returned sequences of constant d18Oc values in individual modeling windows. These issues were caused by a combination of issues in the run_model and cumulative_day functions:

  1. The boundaries on the growth rate sinusoid were set too liberally, causing growth rates to vary by several orders of magnitude within years and allowing growth rate variability over time to be nearly flat or vertical. The run_model function was modified to make the growth rate boundaries more restrictive, and dependent on the estimated initial value for the mean growth rate. The new boundaries still allow growth rate to vary by a factor 10 within a year, which should cover the natural variability of growth rates in d18Oc archives.
  2. Constant d18Oc solutions in modeling windows were often a result of the number of years of d18Oc that were modeled within a window. If the window contained more time than expected the samples outside the d18Oc curve were assigned a constant value, resulting in a flat d18Oc profile. This issue also explained why the mean modeled d18Oc values underestimated the true d18Oc variability in the record.
  3. Slight modifications were made to the cumulative_day function that aligns results from individual modeling windows to a common time axis. The previous version of ShellChron added too many or too little days to the age results in modeling windows in which either none or multiple year transitions occur. This resulted in sudden jumps in time in the age-distance relationship.

0.4.0

Added citation data of the now-published paper Slightly modified d18O_model.r to update the function for VPDB to VSMOW conversion