Master NetCDF Programming with Complete, Tested Examples
Written by Edward Hartnett, co-author of NetCDF, this handbook teaches you NetCDF programming through complete, runnable code examples in C, Fortran 90, and Java. NetCDF (Network Common Data Form) is the industry standard for storing and sharing scientific data across climate science, oceanography, atmospheric research, and high-performance computing.
Why This Book?
Official NetCDF documentation excels as an API reference but can be challenging for learning. This handbook provides a progressive, hands-on approach where every concept is demonstrated through working code you can compile and execute immediately. Unlike tutorials that show fragments, every program here is complete, tested, and ready to run.
What You'll Learn:
- Create and read NetCDF files in Classic (CDF-1, CDF-2, CDF-5), NetCDF-4/HDF5, and ncZarr formats
- Define dimensions, variables, and attributes following CF conventions
- Program in C, Fortran 90, and Java with complete API examples
- Optimize performance with chunking strategies, compression filters, and cache tuning
- Implement parallel I/O using MPI for high-performance computing
- Use command-line tools: ncdump, ncgen, nccopy, and NCO utilities
- Apply best practices from 30+ years of NetCDF development
Book Structure:
Foundation (Chapters 1-5): NetCDF data models, binary format variants, installation, and command-line utilities
Programming APIs (Chapters 6-8): Complete C API, Fortran 90 API, and Java API with NetCDF-Java library and NcML aggregation
Advanced Features (Chapters 9-11): CF conventions, NetCDF-4/HDF5 performance optimization, and parallel I/O with MPI
Production-Ready Code:
All examples include CMake and Autotools build configurations, GitHub Actions CI/CD integration, cross-platform support (Linux, macOS, Windows), and automated test suite. Complete source code available in open-source repository.
Who This Book Is For:
- Climate scientists and researchers managing model output and observational data
- HPC developers implementing efficient parallel I/O for large-scale simulations
- Data engineers building scientific data management pipelines
- Software developers integrating NetCDF into applications
- Students and educators learning NetCDF programming
Prerequisites: Basic programming knowledge in C, Fortran, or Java. No prior NetCDF experience required.
Technical Coverage:
- NetCDF Classic and Enhanced data models
- Multiple binary formats: CDF-1, CDF-2, CDF-5, NetCDF-4/HDF5, ncZarr
- Coordinate variables and CF conventions
- User-defined types: compound, vlen, enum, opaque
- Groups and hierarchical data organization
- Chunking strategies aligned with access patterns
- Compression filters: deflate, shuffle, szip, zstd
- Fill values and missing data handling
- Time encoding with multiple calendar systems
- Parallel I/O: collective vs. independent modes
- Domain decomposition for distributed computing
- NetCDF-Java NcML for virtual datasets and aggregation
