Background In order to accurately determine the linear regression relationship between leaf temperature difference and atmospheric vapor pressure difference (VPD) of Quercus variabilis seedlings under the condition of adequate water supply, and to optimize the crop water stress index(CWSI)empirical model established by predecessors, Zhengzhou city in Henan province, where plant drought is a significant problem, was selected as the main research area, aiming to analyze and understand the dynamic change of soil water and the response degree of plants in a more representative way, so as to build a more applicable CWSI model and soil water diagnosis model.

Methods The temperature, solar radiation, leaf temperature difference and atmospheric vapor pressure difference of Q. variabilis seedlings in the pot experiment were measured by dry-wet reference surface method in research base of College of Forestry, Henan Agricultural University. The reasonable upper and lower baselines of CWSI model of Q. variabilis seedlings were determined by theoretical analysis and calculation of energy balance principle.

Results Under dry reference plane condition, the leaf temperature difference Δt_干 of Q. variabilis seedlings was significantly positively correlated with solar radiation Q, and the upper baseline of CWSI empirical model was determined as a straight line: Δt_干=0.007Q+1.621. Under the wet reference plane, there was a negative correlation between leaf temperature difference Δt_湿 and V(VPD) of atmospheric saturated water vapor pressure difference, and the linear regression relationship was significant. The optimized lower baseline of CWSI empirical model was obtained: Δt_湿=-1.218V+1.987. Thus, the CWSI empirical model after optimizing the lower baseline is I= \fract_\mathrmca-(-1.218 V+1.987)(0.007 Q+1.621)-(-1.218 V+1.987), t_ca is the measured leaf-air temperature difference. The CWSI value of Q. variabilis seedlings was calculated by using the optimized CWSI model with the obtained upper and lower baseline. The linear relationship between CWSI value I and soil moisture θ was significant. The diagnostic formula is θ=-24.65I+27.91. The relationship between CWSI and soil moisture content was as follows: 12% soil moisture corresponded to mild drought, and CWSI was about 0.65. The soil moisture of 8%-10% corresponded to moderate drought, and the CWSI was 0.75-0.8. Soil moisture of 5%-8% corresponds to severe drought and CWSI is 0.8-0.95. When soil moisture was lower than 5% and CWSI was greater than 0.95, the seedling stress was fatal.

Conclusions The relationship between leaf temperature difference and VPD can be more accurately determined by the heat budget calculation method of energy balance equation, so as to optimize the CWSI model, and then calculate the linear relationship between CWSI of different crops and soil moisture according to this model, and obtain the soil moisture diagnosis model. The soil moisture of different crops or forest plants can be diagnosed according to the solar radiation and air temperature and humidity of remote sensing leaves and weather stations.