干旱区生态环境知识资源中心

Hydrothermal time models are used to predict seed germination overtime in seedling establishment models. Soil matric potential data from three resistance-type sensors (Colman cells [Soiltest, Inc. Lakebluff Illinois, USA], Watermark brand sensors [Irrometer Co., Riverside, California, USA], and Delmhorst gypsum blocks [Delmhorst Instrument Company, Towaco, New Jersey, USA]), from a time-domain reflectometry (TDR) probe (Campbell Scientific 615 [Campbell Scientific, Inc., Logan, Utah, USA]), and from gravimetric sampling of the seed zone were input into a hydrothermal time model to determine if we could accurately predict seedling emergence over time in a growth chamber experiment for six desert grass species. Seeds were sown in a structureless sandy loam soil which received four different irrigation amounts and was allowed to dry in a growth chamber programmed to simulate spring soil temperatures in Provo, Utah. Cells and gypsum blocks were more sensitive to soil surface drying than Watermark sensors and TDR probes. Only gravimetric sampling correctly predicted no emergence for the driest soil and fastest drying conditions when surface soil dried out 2-4 days sooner than indicated by the most sensitive sensors. For wetter and slower-drying soils, soil water potential inputs from all sensors were similar and high and predicted seedling emergence within 1-2.5 days of observed emergence. For these wet seed-zone conditions, use of thermal time to predict emergence was as accurate as use of hydrothermal time, evidently because all seeds had high water potentials during germination. Because the soil water sensors we used must be buried deeper than the seed zone, accuracy of hydrothermal time model prediction of field germination using these sensors depends greatly on how fast soils dry out in the sensor measurement zone compared to how fast the seed zone dries out relative to the speed at which seeds accumulate thermal time to germinate.