3.4.3. Air Temperature and Land Use

To obtain the air temperature on 2 m high above land surface (Tasl) conversion of the land surface temperature (LST) using a linear regression model for two, taking into consideration the days used. The regression equations for 12 July 2010 (1) and 11 August 2015 (2) are:

$$\text{Tasl} = 0.20234 \times \text{LST} + 24.22906 \tag{1}$$

$$\text{Tasl} = 0.3278714 \times \text{LST} + 11.68150. \tag{2}$$

The coefficient of determination (*R*2) was 0.65 (for probability level *p* = 0.008467) and 0.68 (for probability level *p* = 0.01112), respectively, for the first and second day. It should be emphasised that the regression models used are only an attempt to quantify the dependencies (Tasl and LST) according to the limited possibilities of their verification, due to difficulties in obtaining the subsequent observations (a limited number of satellite images).

On 12 July 2010 at 9:34 UTC there was a significant diversification of the thermal conditions in the surveyed area (Figure 7a). The spatial distribution of air temperature clearly showed the impact of land development and use. The lowest mean air temperature was 29.8 ◦C over water bodies, and the highest 32.6 ◦C, within the industrial and commercial, public, military and private unit areas (Figure 8a). Among the warmest areas in the city were those with continuous urban fabric, industrial and commercial, public, military and private unit areas, fast transit roads and associated land, other roads and associated land, while among the coldest areas were water bodies, herbaceous vegetation associations and forests (Figure 8a).

**Figure 7.** Air temperature in Pozna ´n on 12 July 2010 (**a**) and 11 August 2015 (**b**) at 9:34. Profile lines (**a**) used in Figure 9.

**Figure 8.** Statistic of air temperature in Pozna ´n ((**a**)—12 July 2010, (**b**)—11 August 2015) on the basis of Landsat images according to Urban Atlas 2012 types (colours and order of types according to legend in Figure 1). In the boxplot, the middle values denote medians; the box extends to the Q1 (first quartile) and Q3 (third quartile), while the whiskers show the range (99.3%). The upper whisker shows Q3 + 1.5 × IQR (the interquartile range) and the lower shows Q1 − 1.5 × IQR. The notches extend to +/−1.58 IQR/sqrt(n) and the dots represent outliers.

On 11 August 2015 at 9:34 UTC, there was less variation in air temperature than on 12 July 2010 (Figure 7b). The impact of land development and use on the thermal conditions was significantly weaker. The lowest mean value of air temperature was 29.7 ◦C over water bodies and the highest, 31.2 ◦C, in continuous urban fabric, airports and mineral extraction and dump sites (Figure 8b).

The effect of land use on thermal conditions is evident in the air temperature profile from the northwest of the city to the southeast (about 22 km) (Figure 9). Along the profile, the air temperature varied from 28.9 ◦C to 33.2 ◦C. The coldest areas were recorded in the northwestern part of the city, i.e., in the western green wedge area, where Kierskie Lake is situated; then there is a systematic increase in air temperature with a maximum in the city centre (>32.0 ◦C), where a slightly colder area of the Warta River is also noted.

**Figure 9.** Air temperature in NW–SE and N–S profiles on 12 July 2010 at 9:43 UTC.

A significant decrease in the air temperature is observed in forest areas in the eastern part of Pozna ´n, forming the eastern green wedge. In turn, along the profile made along the north–south line (about 19 km length), the air temperature changed from 29.6 ◦C to 33.7 ◦C (Figure 9). The coldest areas were found in the northern part of the city, where farmland dominates. A significant increase in air temperature (>32.0 ◦C) was found in areas with low-density, compact urban development occurring in the Naramowice, Winogrady, Stare Miasto and Wilda housing estates. A clear drop in air temperature was recorded in the Warta River Valley.
