The Impact Of The Soil Water Vertical Movement On The Soil Temperature

Soil temperature plays an important part in the study of the interaction betweenthe atmosphere and the land surface. There are many methods to estimate soiltemperature. Two typical methods are introduced in this article: (1) thermalconduction which assumes that soil takes into account thermal conduction alone. (2)The method which couples thermal conduction and convection transfer. The formercontains phase method and amplitude method. We compare these three methods useddata of three different underlying surfaces. Giving the initial temperature profileapproximated by exponential function, we give the solution of the equation ofone-dimensional thermal conduction-convection by the method of Fouriertransformation.Based on soil temperature data collected at sandy desert belt situated in the NWof Negev, Israel in September 30, 2004, we found that the method of Gao gives arealistic estimate of soil temperature while the phase method overestimates thediurnal amplitude by 1.4 K, and the result estimated by the amptitude method showsthe phase shifts by 0.Sh for a soil layer depth of 0.015 m. Meanwhile, using theformer approach, we determine the soil thermal diffusivity and the water flux densityfor four soil layer depths. Results show that both of them change with increasingdepth, which means that the soil thermal properties are heterogeneous in verticalcolumn at this site.Based on soil temperature data collected at a bare soil site over the LoessPlateau during the period from DOYs 197 to 241, 2005, we found that the method ofGao gives a realistic estimate of soil temperature while the phases method, onaverage, overestimates the diurnal amplitude by 0.95 K, and the result estimated bythe amptitude method shows the phase shifts by 0.207 rad (i.e., 47.44 minutes) for asoil layer depth of 0.10 m. The covariance between soil temperature estimated by themethod of Gao, phase method and amplitude method and the observed soiltemperature is 11.58, 12.02 and 11.57 for a soil layer depth of 0.10 m. Using theformer method, we not only determine the soil thermal diffusivity and water flux density, for the soil layer mentioned above, but also for the soil layer depth from 0.10m to 0.20 m, and from 0.20 m to 0.40 m, respectively. We also model the soiltemperature for the depths of 0.20 m and 0.40 m, and the results are in satisfactoryagreement with direct measurements.Based on the soil temperature data collected at the soil site covered withbroomcorn millet site over the Loess plateau during the period from July 23 to31, and August 21 to 29, 2005,, we calculate the thermal diffusivity and the liquidwater flux density with the method of coupled thermal conduction and convectiontransfer at the depths 0.05m to 0.10m in this area. The temperatures are modeled withthe method of coupled thermal conduction and convection transfer, and also modeledwith the method of thermal conduction alone. We compare these two methods above.The covariance between soil temperature estimated by the method of Gao,phasemethod and amplitude method and the observed soil temperature is 8.80, 8.37 and8.36 for a soil layer depth of 0.10 m. The comparison shows the latter overestimatesin day and underestimates in night, because it does not consider the soil watervertical movement, while the former makes a reasonable estimate of soiltemperature.Up-to-date force-restore method assumes that soil is vertically homogenous and.takes into account thermal conduction alone. The present study mathematicallyextends the force-restore method by considering vertical heterogeneity in soil and byincluding thermal convection (e.g. heat transfer by water flux). The averagetemperatures modeled by using the traditional and revised methods for the soil layerranging from 0.035 m depth to 0.10 m depth were compared against the soiltemperature interpolated at 0.0675 m depth from soil temperatures measured at 0.05m and 0.10 m depths at a bare soil site over the Loess Plateau during the period fromJuly 22 to 26, 2005. We compare the correlation coefficient, standard error,normalized standard error between them. Results show that the revised method givesa realistic estimate of soil temperature.