降水波动对呼伦贝尔草甸草原初级生产力年际动态影响

Effects of the Annual Precipitation Fluctuation on Primary Productivity in Hulunbeier Meadow-Steppe

降水是控制草原植被生产力最关键的因素.为探明降水量在不同时间尺度上的波动及其对草原植被的响应机制,以呼伦贝尔草甸草原为研究对象,选用反映年际降水量波动的降水集中度和偏离期2个因子,将以地面光谱生物量模型计算获得的草甸草原植被NPP（net primary production,净初级生产力）与不同周期降水量和降水波动因子建立回归模型,分析年际降水量波动对草甸草原植被NPP的影响.结果表明：1在呼伦贝尔草甸草原区,以年内4—7月为关键期,Pk（关键期累积降水量）对草甸草原植被NPP的影响最大,Cdk（关键期降水集中度）平均值为0.439±0.182,dk（关键期降水偏离期）平均值为31.6 d,变幅为-3.6~94.2 d.2以Pk和以旬为单位的Cdk、dk构建的草甸草原植被NPP估算模型,y=-52.11＋88.957Cdk＋0.724dk＋0.953Pk,能较好地反映草甸草原植被NPP与降水波动之间的关系,模型估测精度可达91.0%.因此,在半干旱草原区,利用基于遥感植被反射光谱构建的NPP模型计算草甸草原植被NPP具有较高的可信度,并且与样方调查结果有极高的关联性.

Rainfall is regarded as one of the important factors affecting productivity of grassland vegetation. To clarify the fluctuation of rainfall on different time scales and its consequential mechanism on steppe vegetation, a Hulunbeier meadow grassland was selected as a study object. Two parameters, rainfall concentration degree and deviated period, which can reflect the variations of rainfall annually, were weighed. A regression model was established based on net primary production （NPP）, which was obtained from aboveground biomass spectral model, different periods of rainfall and its fluctuation factor. Subsequently, the effects of the annual rainfall variation on the primary productivity of the steppe vegetation were also analyzed. The results indicated that ： 1 ） The period from April to July was the key factor, due to the great impact of accumulated precipitation on the meadow grassland primary productivity on Hulunbeier meadow grassland. The mean value of concentration degree （ Cdk ） was 0. 439 ±0. 182. Deviated period （dk ） was 31.6 d, with a range from -3.6 to 94.2 days. 2） The prediction model of the meadow grassland productivity was set by precipitation amount during the key time, together with Cdk and dk , as y = -52. 11 ＋ 88.957Cd~ ＋ 0.724d~ ~ 0.953Pk. It reflects the relationships between the NPP and the variance of precipitation amount with a prediction accuracy of gl. 0%. Hence, the vegetation productivity model from remote sensing data was shown to have higher reliability, and its relevance to quadrat data was verified.