A land data assimilation system (LDAS) based dataset for regional agro-climatic assessments

Xing Liu, Purdue University

Abstract

This study is part of a USDA sponsored project ----Useful to Usable (U2U): "Transforming Climate Variability and Change Information for Cereal Crop Producers". The broader objective includes improving farm resilience and profitability in the U.S. Corn Belt region by transforming existing climate/weather data into usable knowledge and tools for the agricultural community. The specific tasks of this research are: (1) Build a high-resolution (4 km, daily) agro-climatic dataset using a Land Data Assimilation System (LDAS). (2) Estimate regional corn yield across the Corn Belt with crop models and the agro-climatic dataset. (3) Evaluate the impacts of climate variability due to El Niño-Southern Oscillation (ENSO) on corn yield in the Corn Belt. Accordingly, a high-resolution (4 km, 1979-2012, daily) agro-climatic dataset across the U.S. Corn Belt has been built using the North America Land Data Assimilation System version 2 (NLDAS2) product. This newly developed dataset includes daily maximum/minimum temperature, precipitation, solar radiation, soil moisture, and soil temperature at four soil depths (0-10 cm, 10-40 cm, 40-100 cm, and 100-200 cm). Validations indicate strong agreement between this dataset and field measurements. The agro-climatic dataset was then used with a Hybrid-Maize crop model to estimate regional corn yield at grid scale. The crop model was first validated at the field and county scale and found to consistently overestimate yields at the county scale. This was attributed to the optimum field conditions considered in the model and the overall uncertainties. Comparison with NASS yield survey data indicates a 0.6 multiplicative factor provides good agreement with actual yields, and is recommended for county-scale simulations. Following the field/county scale model tests, a modeling framework was developed to simulate gridded crop yields. Results indicate that integrating spatial climatic information improved the regional performance of the Hybrid Maize model and this agro-climatic dataset shows good potential for developing agro-meteorological related applications. Finally, the impacts of the El Nino-Southern Oscillation (ENSO) on observed and simulated corn yields were examined. As a result, La Niña shows a significant negative impact on corn yield in the Corn Belt while the impact from El Niño is insignificant. It also has been found that La Niña correlates with relatively late planting dates in the Corn Belt. Based on a crop model study, the results indicate that for some counties, under optimal conditions, late planting dates can mitigate the negative impacts from the La Niña phase. Based on the studies above, reliable performance of the Hybrid Maize crop model and superior data ability of the new agro-climatic dataset have good potential to simulate regional corn yield with climate projections. The significant impacts of ENSO on corn yield indicate that advance ENSO warning may benefit field management in the Corn Belt.

Degree

M.S.

Advisors

NIYOGI, Purdue University.

Subject Area

Agricultural engineering|Remote sensing

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS