NCCS Hosts GISS ModelE Simulations Assessing the Impacts of Irrigation on Climate


An agriculture crop irrigation system sprays water at sunset in Trempealeau, Wisconsin. Photo by Tony Webster;
acquired from Wikimedia Commons.

As agriculture intensified globally during the 20th century, irrigation rates increased across many regions. The increase in irrigation has been especially rapid since the 1950s with extraction of groundwater resources becoming a common practice.

Despite such significant growth, the distinct effects of irrigation on the Earth’s climate have been the subject of only a few computer modeling studies. For a recently published study, Washington State University, New York University, and NASA Goddard Institute for Space Studies (GISS) researchers ran multi-decade simulations to identify and understand the interactions between simple land-cover change and more complex land-management, specifically irrigation, across nine intensively cultivated regions around the globe.

The NASA Center for Climate Simulation (NCCS) Discover supercomputer hosted three 70-year simulations using GISS ModelE2 running at 2° by 2.5° (the equivalent of 221 km by 278 km at the equator) spatial resolution with 40 vertical levels. Each of the three simulations ran on 112 computing cores, which completed each run in less than 8 hours.

Discover’s online disk stored input data including vegetation and crop cover information and key plant growth attributes taken or adapted from NASA MODIS Land Cover and Leaf Area Index products. The online disk and later the NCCS mass storage archive stored simulation output data of approximately 450 gigabytes.

The three ModelE2 simulations used progressively complex global land surface conditions to serve as sensitivity experiments:

  • NatVeg included only natural vegetation cover and served as a baseline for comparing land surface changes.
  • Crops added modern‐day crop cover from year 2000 estimates for evaluating the impacts of land cover change alone.
  • IrrigatedCrops added year 2000 irrigation rates to the crop cover data for evaluating the impacts of irrigation on agricultural areas.

These experiments led to four main conclusions:

  1. The effects of irrigation on surface temperature and precipitation are comparable or larger than the effects of land-cover changes.
  2. Over some regions irrigation amplifies the effects of land-cover changes (e.g., precipitation over parts of the Mediterranean and West Asia), but in other regions it either dampens or reverses them (e.g., temperature and precipitation changes over parts of South Asia).
  3. Irrigation impacts are not restricted to the primary growing season or peak irrigation season and can vary by season.
  4. The impacts and effects of land-cover change and irrigation are not limited to the surface but extend up to the lower atmosphere, which influences low-level circulation and lower- and upper-level air temperatures and can consequently affect climate in remote regions.

These results highlight the importance of including land management decisions in climate simulations for a more accurate understanding of how human activities shape climate, particularly over the studied regions of intense agriculture. Future scenario-based experiments could help policy-makers consider the climate ramifications of land management and water use decisions.

Simulated changes in (a–c) total annual precipitation and (d–f) annual mean surface temperature in the agricultural land cover change experiment relative to the natural vegetation experiment (Crops − Natural Vegetation), in agricultural land cover change with added irrigation experiment relative to the natural vegetation experiment (IrrigatedCrops − Natural Vegetation), and in the agricultural land cover change with added irrigation relative to the land cover change only experiment (IrrigatedCrops − Crops). Stippling (dots) indicates that changes are significant at the 5% level. Visualizations from Singh, D., S.P. McDermid, et al.

“The NASA GISS climate model is among the few (if only) climate models to incorporate time-varying irrigation,” said Sonali McDermid, assistant professor of Environmental Studies at New York University. “Thus, the model, along with dedicated NCCS computing resources, enables us to efficiently run, post-process, and store model output from novel irrigation-forced climate experiments, all of which cannot be done with the same expediency, efficiency, cost-effectiveness, and support elsewhere.”

“Without a doubt, NCCS allows us to undertake such novel experimentation that I imagine is only feasible in few places around the world, and as such, it is a unique and highly valuable resource for conducting new and cutting-edge research,” she said.

McDermid co-led the study with Deepti Singh, assistant professor in the School of the Environment at Washington State University - Vancouver. Study collaborators were NASA GISS researchers Benjamin Cook, Michael Puma, Larissa Nazarenko, and Maxwell Kelly.

More Information:
Singh, D., S.P. McDermid, B.I. Cook, M.J. Puma, L. Nazarenko, and M. Kelley, 2018: Distinct Influences of Land-Cover and Land-Management on Seasonal Climate. J. Geophys. Res. Atmos., 123, no. 21, 12017–12039, doi:10.1029/2018JD028874.