| A series of research experiments have been conducted in this paper to solve the problem of how to improve the utilization of microclimate resources.Firstly,according to the weather data from the Baodi District of Tianjin during the winter of 2011-2015,the simulation results of different greenhouse temperature simulation methods were compared and analyzed to establish a more accurate simulation model of solar greenhouse environment.Then based on the meteorological data and crop development data of 2013-2015 Sunlight Greenhouse in Wuqing District of Tianjin City,with tomato varieties “Suifen 882” and “Provence” planted in a large area of ? ? facility horticulture were used as the test materials,different meteorological conditions were simulated through staged sowing experiments,and period simulation model of tomato development in the solar greenhouse was established based on the bell model.Combining the simulation results of the above two studies,a simulated analysis of the time-to-market of solar greenhouse vegetables in Baodi District,Tianjin City during the winter of 2016-2017 was conducted to provide decision-making guidance for actual planting time and management,and to achieve efficient use of greenhouse microclimate resources.The main findings are as follows:1.Environmental simulation,in order to build a more accurate solar greenhouse temperature forecasting model,environmental monitoring experiments were conducted inside and outside the greenhouse during the winter of 2011-2014(January,February and December)in Baodi District,Tianjin,a prediction model based on stepwise regression and BP neural network greenhouse temperature of three weather types(Clear,cloudy,cloudy)during three periods(0-8 hours,8-17 hours,17-23 hours)were established.The results show that within the nine conditions of stepwise regression inner greenhouse temperature model,the average accuracy with the difference between the simulated and actual values less than 3°C is 88%,and the average root mean square error(RMSE)is 2°C.While for the BP neural network model,the average accuracy with the difference between the simulated and actual values less than 3°C is 94%,and the average root mean square error(RMSE)is 1.6°C.The temperature prediction model established by BP neural network is relatively more accurate and stable.2.Crop simulation,accurately predicting the development period of crops based on environmental conditions has become one of the core contents of crop growth and development simulation models.Based on the clock model for the first time,followed by the light-temperature response characteristics of tomato growth and development,and using thetest data of 12 growing seasons for different cultivars at different sowing dates,a simulation model of greenhouse tomato development period based on the bell model method was established.The regression between the simulated and observed values ? ? during the five stages of sowing,three-leaf stage,sowing stage,early flowering stage,sowing stage,fruit setting stage,sowing stage,maturity stage,sowing stage and pulling stage was verified.The estimated root mean square errors(RMSE)were 8.3,14.4,16.3,23.7,and 28.1 days,respectively;the regression estimates of the mean square root error(NRMSE)were 20.78%,20.18%,20.21%,17.35%,and 14.86%,respectively.better.Comparing this model with the simulated results of effective accumulative temperature method,this model has higher simulation accuracy and better results in each key development period.3.Practical application,based on the 58-year meteorological data recorded in Baodi District of Tianjin,calculate the annual mean temperature of the 58 th year,the annual average temperature anomaly,the average annual temperature of 5 years,and the average number of hours of sunshine.The annual average of sunshine hours,annual sunshine hours,annual sunshine hours,and five-year moving averages,and studied the trends of temperature and light changes in 58 years.The results of the analysis show that the winter temperature in Baodi District has significantly increased in the past 58 years,and the trend of changes in the overall illumination has weakened.It has been further concluded that the cold and warm year division method considering the temperature and light conditions is more suitable for the production of solar greenhouses.In this study,the meteorological factors of average temperature and sunshine hours were selected to obtain the calculation equation of the division standard of cold and warm years: y=0.608xt+0.392 xr.The y value was used as a standard to divide the cold and warm year types suitable for greenhouse production,and the results were divided into 2016.Warm year for the greenhouse.Selecting the BP neural network method with higher accuracy to simulate the daily average daylight temperature and daytime sunshine hours of the solar greenhouse in winter2016-2017,and obtaining the daytime daily average temperature and the daytime sunshine hours crop driving factors brought into the tomato development model,according to After the different sowing dates are drawn,the development period is changed.This paper will use the actual situation of large quantities of vegetables in the early spring festival as an example to guide the planning of tomato planting time.The best sowing date for the tomatoes to be listed before the Spring Festival of 2017 is September 20,2016-September 25,2016. |
PDF Full Text Request