*2.3. Seasonal Characteristics*

Firstly, the data of a typical day in spring and summer was selected, respectively. Then, the time-lag curve of temperature and temperature-induced strain was drawn, see Figure 3. It is obvious that the temperature and response strain curve possess typical seasonal characteristics. By comparing Figure 3a,b, it is observed that the hysteresis loop in spring, as shown in Figure 3a, is relatively more compact than that in summer, as shown in Figure 3b. It reflects that in spring the change in temperature is comparatively slower. Hence, the degree of strain response lagging temperature is also less than that in summer. It also gives us a new way to measure the extent of time-lag e ffect through the hysteresis loop area.

**Figure 3.** Plots of strain vs. temperature.

Furthermore, to illustrate the time-lag e ffect as a common phenomenon of concrete box-girder bridges, this paper investigated three months of data. Additionally, the hysteresis loop area of temperature and strain data of every single day are plotted in Figure 4. From Figure 4, a time-lag phenomenon of di fferent extent in almost every day indeed exists. Moreover, the hysteresis loop area in summer (July) is larger than that in winter (November).

In summary, this part gives us a clear understanding of the time-lag phenomenon of temperature effects, a very common phenomenon in concrete box-girder bridges. Moreover, this phenomenon in the form curve annular feature can be directly reflected by the temperature vs. strain graph. Based on the analysis of data in di fferent seasons, a seasonal characteristic was found.

**Figure 4.** Hysteresis loop area in three months of 2017.
