BiostatR Blog

ERA5 is the latest development in the ERA series, and  improves significantly on its predecessors by: - Offering a higher horizontal resolution of 31 km and 137 vertical levels from the surface up to 0.01 hPa (around 80 km); - Using a more recent and advanced version of the ECMWF IFS model; - Providing hourly estimates of atmospheric variables; - Providing a consistent representation of uncertainties for these variables; - Using more satellite observations in the data assimilation. Here are some information to access this database using python batch Please visit http://climate.copernicus.eu/climate-reanalysis for more information including documentation and guidelines on how to download the data. __________ June 2025 The new ECMWF methodology cannot use R directly. The solution I found to download data in MacOSX was: https://cds.climate.copernicus.eu/datasets/reanalysis-era5-single-levels?tab=overview To install the CDSAPI using pip, the solution is to create a virtual environneme...

 Error because library libintl.h is not found; to solve: sudo ln -s /opt/homebrew/Cellar/gettext/0.25/include/libintl.* /opt/R/arm64/include

 atT <- structure(c(1672520400, 1675198800, 1677618000, 1680296400, 1682888400,              1685566800, 1688158800, 1690837200, 1693515600, 1696107600, 1698786000,              1701378000, 1704056400, 1706734800, 1709240400, 1711918800, 1714510800,              1717189200, 1719781200, 1722459600, 1725138000, 1727730000, 1730408400,              1733000400, 1735678800), class = c("POSIXct", "POSIXt"), tzone = "Asia/Riyadh") # Work as expected plot(x = atT, y=rep(1, length(atT)), xaxt="n") axis(side = 1, at=atT, labels = as.character(seq_along(atT)))    # Only first label is shown plot(x = atT, y=rep(1, length(atT)), xaxt="n") axis(1, at=atT, label=c("Jan-2023", "", "", "", "", "", "Jul-2023", "", "", "", "", "",                                  ...

Let X, one observation composed of two values (x1, x2) obtained from Gaussian distributions µ1, s1 and µ2, s2 with rho being a correlation coefficient between Gaussian distributions 1 and 2, the likelihood is: library(mvtnorm) dmvnorm(x=c(x1,x2), mean=c(µ1,µ2), sigma=matrix(c(s1, rho, rho, s2), ncol=2))

 These options seem to be ok: ```{r setup, include=TRUE} knitr::opts_chunk$set(fig.path='figs/xxx_',                        dev='svg',                        concordance=TRUE,                        echo=TRUE,                        fig.width=12, fig.height=8,                        dev.args=list(pointsize=16)) ```

 MacOSX 15.3 gcc installed using MacBrew I get this error: ld: warning: search path '/opt/gfortran/lib/gcc/aarch64-apple-darwin20.0/12.2.0' not found ld: warning: search path '/opt/gfortran/lib' not found ld: library 'gfortran' not found I solve it by creating a  ~/.R/Makevars  file like this: FC      = /opt/homebrew/opt/gfortran/bin/gfortran F77     = /opt/homebrew/opt/gfortran/bin/gfortran FLIBS   = -L/opt/homebrew/opt/gfortran/lib Now I can compile package with gfortran