---

作 者 ：孙成明,庄恒扬,杨连新,杨洪建,黄建晔,董桂春,朱建国,王余龙

期 刊 ：生态学报 2007年 27卷 2期 页码：613~619

关键词：水稻；FACE；生育期；模拟模型；

Keywords:rice, free-air CO2 enrichment, growth duration, simulation model,

摘 要 ：用农田开放式空气CO₂浓度增高（FACE）技术平台，设置大气CO₂浓度比对照高200μmol mol^-1的FACE处理和不同的施N量水平，以水稻钟模型为基础，构建了FACE水稻生育期模拟模型。通过不同年度试验数据的检验，表明模型对CK及FACE条件下水稻不同生育期天数的预测性能好。不同生育期预测误差的根均方差（RMSE）最大为2.64d，最小只有0.15d，且相关系数均达到了极显著水平。说明FACE水稻生育期模型具有较高的预测性和适用性。

Abstract:Global atmospheric CO₂ concentration([CO₂]) was projected to reach levels of 650-780 μmol mol-1 at the end of this century(IPCC,2000). The increasing atmospheric [CO₂] and associated climate may greatly affect agricultural productivity worldwide. However, little was known about the effect of free-air CO₂ enrichment(FACE) on growth duration of rice. In order to predict the changes of growth duration of rice under FACE condition, we conducted a FACE experiment at Anzhen(31°37′N,120°28′E), Wuxi, Jiangsu province, in 2001-2003.
The FACE facility consisted of three FACE rings(three repetitions) and five ambient rings, and the ambient rings were 90m apart to avoid contamination of elevated CO₂. Rice were grown from transplanting to harvest under two levels of [CO₂] for three consecutive years(2001-2003), and the target [CO₂] of the FACE rings was 200μmol mol-1 above that in the ambient air. The annual mean temperature was about 16℃, the annual mean rainfall was 1150 mm and the annual sunshine hours were 2000h in the experiment area.
The experiment was conducted in 3 years using the japonica cultivar Wuxiangjing 14. Seeds were sowed on 18 May and seedlings were hand transplanted with 3 plants each hill on 13 June. Hills and rows were 16.7 and 30 cm apart, respectively(equivalent to 24 hill m^-2). Three levels of N application were: 150kg/hm²(low,LN), 250kg/hm²(medium,MN) and 350kg/hm²(high,HN). To calculate the days of main growth duration of rice in both ambient and FACE plots,10 hills per experiment plot were selected and marked to investigate leaf age. Results showed that: the mean days from sowing to heading, from heading to maturity and the whole growth duration of rice shortened 3-4.5d(mean value 3.4d), 2-3d(mean value 2.4d) and 5-7.5d(mean value 5.8d) in FACE treatment compared with CK. Increasing of nitrogen application rate could weaken the effect of FACE on growth duration.
In order to describe these changes and predict the growth duration of rice, we developed a growth duration model of rice in FACE condition based on the “rice clock model”. The factor of N contents in rice plants was introduced to adjust the growth duration model in FACE rice. The model was tested with different experimental data and the errors between simulated days and observed days in the CK and FACE conditions were small, with the RMSE in different growth duration from the biggest 2.64d to the smallest 0.15d. The results showed that the growth duration model in FACE rice was not only accurate but also applicable.

---