Modeling On Crop Yield And Soil Water, Heat And Nutrient Dynamics For Different Tillage Managements And Fertilizations

Tillage and fertilization are two factors impact the soil water-hear-nutrientprocesses, long-term unreasonable tillage and fertilization can lead to yieldreduction, loss of water and soil resources and environment pollution. Thus,studying the effect of tillage and fertilization on crop yields, soil water, heat andnutrient dynamics, to develop the best management practice to improve the cropyields and soil situations, maintain the healthy environment. In this study, weused both the field experiment data and Decision Support System forAgrotechonlogy Transfer (DSSAT) Crop System Model (CSM) to analysize thedifferent field managements and to evaluate the simulation performances. Themain results as below:From the year of2009to2011, in Northeast China,3-year filed experimentdata were used to evaluate the long-term effects of no-till (NT), reduced tillage(RT) and conventional tillage (CT) on soil water storage, water use efficiency,and soybean and corn yields. The soil water contents under NT were higher thanCT. And NT had the greatest soil water storage and CT the least. The water useefficiency was greatly affected by weather conditions and tillage managements,i.e., CT and RT were36%and15%higher than NT in2009, respectively. Themeasured corn grain yield data under NT decresed than those under CT, but thegrain yields under RT did not decrease. Thus, without conserdering theenvironmental benefits, we conclude that RT was a good compromise for bothsoybean and corn, as it was associated with higher economic benefits for farmers.From the year of2004to2011, in Northeast China,8-year filed experimentdata were used to evaluate the ability of the DSSAT model to predict crop yields and root zone soil water and temperature dynamics for a soybean-maize rotationunder CT, RT and NT.“Good” agreement between simulated and measuredyields was achieved for model calibration (nRMSE=9%-15%), and “good” to“moderate” agreement was achieved for model evaluation (nRMSE=12%-17%).Simulated volumetric soil water contents in the top20cm layer of CT, RT andNT were in “moderate” to “good” agreement with measurements (d=0.81-0.91;nRMSE=15.3%-20.0%). Overall agreements between measured and simulatedsoil temperature varied from “poor” to “excellent” depending on year and tillage;and the measured soil temperatures were consistently overestimated (E=3.2-6.2℃), possibly due to lack of accounting in DSSAT for the insulating effects ofaccumulated surface residues, and the shading effects of standing crops.Refinement of the soil temperature algorithm in DSSAT is recommended.From the year of2007to2012, on Southwest Ontario, Canada,6-yearcontinuous maize experiment with fertilization (CC-F) and without fertilization(CC-NF) data were used to evaluate the DSSAT model. A good agreement wasachieved in the calibration from the CC-F yield data (d=0.99, nRMSE=23%andE=-29kg ha-1) while a moderate agreement was achieved in the evaluation ofyield data in the CC-NF treatment (d=0.74, nRMSE=43%and E=55kg ha-1).The simulated above-ground biomasses were in good agreement with themeasured data in both treatments (d ranged0.82-0.99) depending on the year andfertilization treatment. The simulated soil inorganic N contents in moderate togood agreements with the measured data in the CC-F treatment (d ranged0.76-0.88and nRMSE ranged35.8%-54.0%), which is better than the CC-NFtreatment (d≤0.5and nRMSE=72.0%-81.0%). The performance of the model onsoil water content were in “moderate” agreements in2007and2008for bothtreatments (d=0.60-0.76). The nRMSE values (ranged25.2%-35.9%) indicated asmaller discrepancy of soil water content compared with the nRMSE values in other evaluated variables. The overall evaluation showed that the DSSAT modelreasonably simulated maize growth, soil mineral N and soil water dynamics inSouthwest Ontario, Canada.