A hypothesis that the increasing application of both surface and ground water for irrigation of crops is a significant source of anthropogenic global warming is tested. In climate models, water is already assigned a major secondary amplifying role in warming, solely as a positive feedback from an atmosphere previously warmed by other GHGs. However, this conclusion ignores the direct anthropogenic forcing from increasing use of water in dry regions to grow crops for the human population. The area irrigated worldwide increased by around 1.5% annually between 1960 and 2000, almost trebling in magnitude. Importantly, though only a small proportion of the Earth’s surface, this additional water vapour is dynamically focussed on dry land, intensifying its potential to elevate the troposphere and reduce the regional OLR. Our modelling analysis suggests that the increase in atmospheric water vapour from irrigation could be significantly more than 1% by 2050 compared to 1950, imposing a global forcing exceeding 1.0 W/m2. Fortunately, this hypothesis can be tested, for example, using the satellite data on OLR acquired since the 1970s, relating this to local trends of increasing irrigation or major floods in arid regions. If found consistent with the data, current proposals to mitigate climate change by limiting combustion of fossil fuels may prove less effective. This prediction regarding the warming effect of increasing irrigation is tested using NCAR reanalysis data made possible by the natural experiments of the periodic flooding of Lake Eyre in Australia's semi-arid interior. It is recommended that this hypothesis be tested using data from local studies in irrigated regions such as changes in outgoing longwave radiation and in increased absorption of incoming shortwave radiation in air.