From: Quantifying how urban landscape heterogeneity affects land surface temperature at multiple scales
Scale | Equation | r | R2 | P value |
---|---|---|---|---|
Sub-region | LST = 43.01 − 0.04 contrast of NDVI | − 0.17 | 3.2% | 0.4 |
 | LST = 44.75 − 0.04 mean of NDVI | − 0.67 | 44.9% | 0.01 |
 | LST = 30.89 + 0.08 energy of NDVI | 0.72 | 52.1% | 0.00 |
 | LST = 50.45 − 0.04 entropy of NDVI | − 0.70 | 49.5% | 0.00 |
 | LST = 42.37 − 0.03 variance of NDVI | − 0.02 | 0% | 0.99 |
 | LST = 35.70 + 0.04 homogeneity of NDVI | 0.69 | 48.1% | 0.01 |
 | LST = 32.19 + 0.04 correlation of NDVI | 0.25 | 6.6% | 0.23 |
 | LST = 44.02 − 0.04 dissimilarity of NDVI | − 0.48 | 23.1% | 0.02 |
 | LST = 14.331 + 3.530 correlation + 4.02 energy − 0.1744 mean | – | 86.08% | 0.00 |
1000m | LST = 45.95 − 0.02556 contrast of NDVI | − 0.08 | 0.7% | 0.40 |
 | LST = 52.60 − 0.4834 mean of NDVI | − 0.51 | 26.9% | 0.00 |
 | LST = 43.75 + 15.00 energy of NDVI | 0.45 | 20.7% | 0.00 |
 | LST= 55.21 − 3.831 entropy of NDVI | − 0.54 | 29.4% | 0.00 |
 | LST = 45.74 − 0.01540 variance of NDVI | 0.03 | 0.2% | 0.68 |
 | LST = 39.67 + 12.55 homogeneity of NDVI | 0.53 | 28.3% | 0.00 |
 | LST = 44.40 + 4.502 correlation of NDVI | 0.17 | 3.1% | 0.00 |
 | LST = 48.08 − 1.151 dissimilarity of NDVI | − 0.35 | 10.4% | 0.00 |
 | LST = 59.69 + 10.91 correlation + 1.481 dissimilarity − 3.940 entropy − 0.7064 mean | – | 62.6% | 0.00 |
Pixel-based | LST = 45.74 + 0.000720 contrast of NDVI | 0.006 | 0.00 | 0.83 |
 | LST = 48.94 − 0.2259 mean of NDVI | − 0.29 | 9% | 0.00 |
 | LST = 44.79 + 8.686 energy of NDVI | 0.19 | 3.8% | 0.00 |
 | LST = 48.91 − 1.256 entropy of NDVI | − 0.18 | 3.3.% | 0.00 |
 | LST = 45.72 + 0.001848 variance of NDVI | 0.01 | 0% | 0.68 |
 | LST = 44.11 + 3.437 homogeneity of NDVI | 0.18 | 3.3% | 0.00 |
 | LST = 45.57 + 0.7066 correlation of NDVI | − 0.06 | 0.4% | 0.10 |
 | LST = 46.05 − 0.1361 dissimilarity of NDVI | − 0.07 | 0.6% | 0.10 |
 | LST = 43.90 + 0.04853 variance − 0.4057 mean + 2.113 entropy + 15.16 energy | – | 17.3% | 0.00 |