The WMO/ETCCDI report “Guidelines on Analysis of extremes in a changing climate in support of informed decisions for adaptation” [Klein Tank et al., 2009] defines growing season length as the annual count of days between first span of at least six days where TG (daily mean temperature) > 5ºC and first span in second half of the year of at least six days where TG < 5ºC: Count the annual (1 Jan to 31 Dec in Northern Hemisphere) number of days between the first occurrence of at least six consecutive days where T G > 5ºC and the first occurrence after 1 July of at least six consecutive days where TG < 5ºC.
Gridded TG data area available to download as NetCDF files for E-OBS Version 14.0. Gridded E-OBS climate indices are also available to download as NetCDF files on a regular 0.25 degree grid. These include Annual Growing Season Length (GSL), as define by ETCCDI. Note that GSL is based on E-OBS Version 11, while TG data are from the latest version (V14). Differences in versions are described for v14 here and for older versions here. Visual comparisons of the plots they have produced don’t show any noticeable differences in mean temperatures over Ireland, but if I find differences between my calculated GSL values (v14) and the ECAD GSL values (v11), I should check version differences more closely.
The Plan
- Compare / associate to Gunnera locations.
- Compare myGSL to GSL calculated from Station Data.
Data
I've downloaded the E-OBS data as NetCDF files and saved them on my UCD computer here:
/Users/conor/DATA/ECAD/Gridded
/Users/conor/DATA/ECAD/Gridded
- tg_0.25deg_reg_v14.0.nc is 4.2G
- gsl_yr_025deg_reg_v11.nc is 23M
myGSL
I've written my code as a python jupyter notebook. Code is saved on my Google Drive:
Research/Bruce/GrowingSeasonGunnera/GS-PythonCode
My code (GrowingSeasonSpatial.ipynb) calculates the Start of Growing Season (SGS), End of Growing Season (EGS) and the season length (myGSL).
An area plot of the year with the largest mean GSL difference, 1988, shows area differences:
Selecting the point with the largest difference in GSL for this year, I checked the values by hand by printing all TG values for that year at that location. My code gives SGS = 1, EGS = 330, myGSL = 330. Checking by hand confirmed these values. The ECAD GSL value at this location is 224. If the temperature on day one were below 5, though, the SGS would be 83 instead of 1, so I can see how small differences in the dataset could produce large changes to the GSL.
The differences could be due to the fact that ECAD GSL is calculated using v11 data, while I am using v14 data. Unfortunately, v11 data are no longer available to download. I have downloaded v12 data and ran my code again with v12 data. The results are still different to the ECAD data, and closely agree with my v14 results.
Interestingly, both my code and the ECAD code contain GSL values > 361, which strictly speaking should not be possible. My code does this as it performs a rolling sum over the whole dataset, before separating by year, therefore the start of the growing season can be on day 1, if the last 5 days of the previous year also had temperatures > 5ºC. I assume the ECAD code is doing something similar, as it has over 1000 GSL values > 361 out of 31200 values.
Research/Bruce/GrowingSeasonGunnera/GS-PythonCode
My code (GrowingSeasonSpatial.ipynb) calculates the Start of Growing Season (SGS), End of Growing Season (EGS) and the season length (myGSL).
Compare myGSL to ECAD GSL
myGSL and ECAD GSL values don't always agree:
The differences could be due to the fact that ECAD GSL is calculated using v11 data, while I am using v14 data. Unfortunately, v11 data are no longer available to download. I have downloaded v12 data and ran my code again with v12 data. The results are still different to the ECAD data, and closely agree with my v14 results.
Interestingly, both my code and the ECAD code contain GSL values > 361, which strictly speaking should not be possible. My code does this as it performs a rolling sum over the whole dataset, before separating by year, therefore the start of the growing season can be on day 1, if the last 5 days of the previous year also had temperatures > 5ºC. I assume the ECAD code is doing something similar, as it has over 1000 GSL values > 361 out of 31200 values.
References
[Klein Tank et al., 2009] Klein Tank, A. M., Zwiers, F. W., and Zhang, X. (2009). Guide- lines on analysis of extremes in a changing climate in support of informed decisions for adaptation. Technical Report WMO-TD No. 1500, World Meteorological Organization.
