Table A4.
In Portugal, the proportion of urban landscape in the area around observation locations differed in April 2020 compared to the prior four years. For each of the four moments of the distributions — mean, variance, skewness, and kurtosis — the coefficient describing the 2020 value is the reference category, and the coefficients for all other years describe deviations from the 2020 reference value. Coefficient and standard error values printed in bold font denote coefficients that were estimated with high statistical precision: the 95% confidence intervals around these coefficients did not overlap zero. Thus, if the coefficients for effects in the years 2016–2019 either all positive or all negative, and all are displayed in bold font, then in 2020 a moment of distribution is consistently different than in all prior years examined. The coefficients for describing mean and variance are presented on the logit scale, while skewness and kurtosis are presented on the ln-link scale.
| Moment | Predictor | Coefficient | SE |
|---|---|---|---|
| Mean | Intercept (2020) | −0.53 | 0.03 |
| Year (2016) | −0.14 | 0.05 | |
| Year (2017) | −0.20 | 0.05 | |
| Year (2018) | −0.29 | 0.04 | |
| Year (2019) | −0.31 | 0.04 | |
| Variance | Intercept (2020) | 0.34 | 0.03 |
| Year (2016) | −0.20 | 0.05 | |
| Year (2017) | −0.21 | 0.04 | |
| Year (2018) | −0.44 | 0.03 | |
| Year (2019) | −0.31 | 0.04 | |
| Skewness | Intercept (2020) | 0.09 | 0.04 |
| Year (2016) | 1.03 | 0.06 | |
| Year (2017) | 0.88 | 0.05 | |
| Year (2018) | 0.24 | 0.05 | |
| Year (2019) | 0.68 | 0.05 | |
| Kurtosis | Intercept (2020) | −1.70 | 0.07 |
| Year (2016) | −0.90 | 0.17 | |
| Year (2017) | −0.67 | 0.12 | |
| Year (2018) | −0.79 | 0.11 | |
| Year (2019) | −0.48 | 0.11 |