NoData

NoData is a sentinel value used in raster data to signify pixels that do not contain valid data. In geographic information systems (GIS), NoData serves to distinguish between measured values and regions that are outside the dataset, masked, or otherwise unavailable. It is not a real measurement; rather, it is a flag indicating missing or inapplicable information. The interpretation of NoData can vary by format and software, but its purpose remains to prevent invalid calculations or misinterpretation of results.

Most raster formats store NoData as a dedicated value in the dataset's metadata. Examples include GeoTIFF, ERDAS IMG, NetCDF, and ASCII grid. The specific numeric sentinel is dataset dependent, with common choices such as -9999. When processing, GIS tools typically treat NoData as transparent or ignored during calculations, visualizations, and resampling. If two rasters with different NoData values are combined, tools generally harmonize to a single NoData reference or explicit masking.

Any analysis that does not account for NoData can produce biased results. Users should verify the NoData value used in each dataset, convert to a consistent NoData convention when merging layers, and understand how the software propagates NoData through algebraic operations. NoData is distinct from zero or NULL in many data systems, and in floating point arrays, some environments use NaN as a missing value; however, NoData in raster data is a formal metadata flag used across many formats.