b'CommentaryData trendsequal to their dimension namesake.The Data trendsdeclaration of Relative Humidity (rh) shows the logical intentions of constructing a \x1coating-point multidimensional array Tim Keepingwith the dimensions. Range limits, long Associate Editor for geophysical data managementnames and units can be set.and analysis Variables are usually numeric, but the Etechnical-standards@aseg.org.au format shows its Fortran and C roots with UTF-8 character arrays instead of strings. Strings were introduced later, as was NetCDFData Form Language (CDL) examplethe Coordinates component allowing constructing an early NetCDF. variables to be accessed through locations. The ASEG GDF2 text format opens on anylat = 5, lon = 10, level = 4, time =Variables with unlimited dimensions can computer, and users can open it in Excelunlimited be appended whereas the others are set. if they need to \x1ex anything, so many areAttributes are optional descriptions such happy that it is the default data exchangeint lat(lat), lon(lon),level(level);and archive format. However, extremelyshort time(time) as the long name orrange.large statewide government airborne\x1boat rh(time,lat,lon); The original point of NetCDF was to surveys usually exceed the 1-million-linerh:long_name = relative humidity; follow NASAs lead and standardise an constraint of Excel, and drone surveyorsopen, portable \x1ele format and its access using 1000 Hz magnetometers have torh:valid_range = 0.0, 1.0;// min andfunctions to reduce time, e\x1fort and severely down sample to be under thatmax errors by users. This aligns with the ASEG magical million line limit. data// optional datageneral ethos. According to the Library So, we (the ASEGs Technical Standardsassignments of Congress assessment NetCDF has few Committee) began investigating NetCDFlevel= 1000, 850, 700, 500; concerns beyond compression (Figure 1).(Net Common Data Form) as a GDF2lat = 20, 30, 40, 50, 60; I downloaded a survey from Geoscience substitute, and it turned out that thislon= -160,-140,-118,-96,-84,-52,-45,Australia GADDS 2.0 as both ASEG GDF2 format is not just set of HDF wrappers-35,-25,-15; and NetCDF \x1eles. They were almost the but was originally wrappers for NASAstime= 12; same size compressed at ~250 MB, but Common Data Format (CDF) for storingwhen expanded the GDF2 grew to 1.6 GB multidimension arrays.rh=.5,.2,.4,.2,.3,.2,.4,.5,.6,.7, whereas the NetCDF stayed unchanged NetCDF was developed and is maintained.1,.3,.1,.1,.1,.1,.5,.7,.8,.8, with built in compression.by Unidata (https://docs.unidata.ucar.edu/etc But what can you do with it if you are not netcdf-c/current/), part of the UniversityThis is a nice, simple to understand exampleinterested in programming to extract Corporation for Atmospheric Researchde\x1ening integer arrays with upper boundsNetCDF data? Not many commercially (UCAR). UCAR is funded principally by the National Science Foundation. The software is currently at version 4.9.2 (https://www.unidata.ucar.edu/software/netcdf/).The McGill University website has a clear and concise history and descriptions https://www.bic.mni.mcgill.ca/users/sean/Docs/netcdf/guide.txn_1.html. The multidimensional aspect of the format means that it can hold multiple grids of the same dimensions, and the open source GDAL is the easiest way to convert grids to NetCDF. This would be useful if you are accumulating too many grid formats, but is it an answer for storing and exchanging large numbers of pointdata?A walk-through example from McGill University demonstrates that it is easy to understand this tool. A NetCDF \x1ele has components - dimensions, variables and data, with optional attributes. Dimensions de\x1ene names with lengths, that can only be integers and will de\x1enethe size and shape of variables. The following is anFigure 1: Summary diagram of NetCDF \x1dle format. https://docs.unidata.ucar.edu/netcdf-c/current/excerpt from the McGill sites Commonnetcdf_data_model.htmlAPRIL 2024PREVIEW 35'