Contribution Analysis Of Driving Factors For Maize Yield Growth In China

China's maize yield has demonstrated a slowing growth since the end of last century.This has received great concerns for policy makers and agricultural scientists.Reasons for such phenomena are usually ascribed to the decline in fertilizer efficiency,land degradation,the reduction in technology inputs such as new crop varieties,and climate change.However,none of the reasons has been profoundly investigated and robustly quantified to date,particularly at national or regional scale.We applied an ensemble empirical mode decomposition(EEMD)analysis on historical statistical data,and discovered that the slowdown in maize yield growth appeared at both the national scale and individual 24 provinces(autonomous regions,municipalities)in 1981-2010.We then conducted statistical and EPIC crop modeling analysises at a gridded level,to untangle the contributions of five factors(climate change,fertilizer,area,irrigation and cultivar)on maize yield growth in 1981—2010.The preliminary conclusions are as follows:(1)There was an increasing trend of maize yield but slowing yield growth had been experienced at the whole China and 24 provinces(autonomous regions,municipalities)in 1981-2010.(2)Results from multiple linear regression showed,at national scale,maize yield was significantly correlated with fertilizer,irrigation,planting area and climate factors during 1981-2010.Maize yield was significantly promoted by inputs of fertilizer,irrigation and planting area.A 1% increase of these investments,respectively lead to a growth of maize yield by 0.27%,0.15% and 0.08%.Among climate factors,enhanced precipitation increased maize yield,with a 1% increase in precipitation promoting maize yield by 0.09%.Whereas temperature and cloud cover exhibited negative effects on maize yield change.A 1% increase in temperature and a 1% decrease in solar radiation would decrease maize yield by 0.87% and 0.35% respectively.Fertilizer addition contributed most to past yield increase of maize at national scale(40.10%).At regional scale,past increase of fertilizer application amount also contributed most to past yield increase of maize in Northern spring maize region(32.83%)and Huang-Huai-hai summer maize region(36.77%).Compared with fertilizer,the contribution of climate change,planting area and irrigation were small.Within all climate drivers,the increase of temperature reduced maize yield by 2.83%,while the decrease in precipitation and solar radiation reduced maize yield by 0.17% and 1.54%.(3)EPIC simulation also confirmed the statistical conclusion that fertilizer addition was responsible for past yield increase of maize.Technology advance such as new crop varieties demonstrated increased contribution to maize yield growth during the past three decades.Climate change and the change of irrigation had positive effects on maize yield change in the whole country and Northern spring maize region,while negative effects in Huang-Huai-hai summer maize region in 1981—2010.The contribution of maize planting area was the lowest among five driving factors.(4)Statistical method and mechaninic method both verified that increased fertilizer application contributed most to past yield increase of maize,while its contribution experienced a declining trends over time,partly resulting in observed slowing increase of maize yield.Estimated contribution for fertilizer from statistical method is lower than that estimated by EPIC.The increase of maize planting area and the decline of irrigation coefficient respectively exhibited positive and negative effects on maize yield under statistical method,which contrast to the estimation by EPIC.This is also true for climate change effect,statistical method showed a negative effect of climate change,which a positive effect with EPIC.EPIC model showed that climate change had positive effects on maize yield change in the whole country and Northern spring maize region,but negative effects in Huang-Huai-hai summer maize region.The EPIC model can research the effect of cultivar on maize yield while statistical method can not quantify the contribution of cultivar on maize growth.The accuracy of the results is affected by the selection of datasets,regression method and model parameterization.Our study mainly focus on the major factors that affect China's maize production,which to some extent has limited representation as many other factors are not considered here,such as environmental degradation,pest,diseases,and labor.Nevertheless,our results are consistent with previous studies,confirming that fertilizer plays a significant role in maize yield growth,and climate change has a growing importance in driving the future production trends.Next work will focus on the uncertainty and sensitivity analysis,where a couple more models will be selected to explore the impacts of more factors on maize production and improve the accuracy of the results.