Wei Chong1,2,*, Wenhua Lü1,2, Jian Zhang1,3,4, Jing Liang1..., Xiaotong Yang1 and Guoyu Zhang1,3,4|Show fewer author(s)
Author Affiliations
1School of Opto-Electronic Engineering, Changchun University of Science and Technology, Changchun, Jilin 130022, China2Meteorological Observation Center of CMA, Beijing 100081, China3Key Laboratory of Opto-Electronic Measurement and Optical Information Transmission Technology, Ministry of Education, Changchun, Jilin 130022, China4Opto-Electronic Measurement and Control Instrumentation, Jilin Province Engineering Research Center, Changchun, Jilin 130022, Chinashow less
Fig. 1. Fitting curves and surfaces of diffuse fraction Sigmoid model. (a) S0 model fitting curve based on 1 h training data; (b) S0 model fitting curve based on 10 min training data; (c) S0 model fitting curve based on 1 min training data; (d) S1 model fitting surface based on 1 h training data; (e) S1 model fitting surface based on 10 min training data; (f) S1 model fitting surface based on 1 min training data
Fig. 2. Validation comparison diagrams for S0 models with different parameters. (a) Validation comparison diagram for S0 model using 1 h radiation data collected in May, 2012 at Lijiang station; (b) validation comparison diagram for S0 model using 10-min radiation data collected in May, 2012 at Lijiang station; (c) validation comparison diagram for S0 model using 1 h radiation data collected in July, 2012 at Kazuo station; (d) validation comparison diagram for S0 model using 10-min radiation data collec
Fig. 3. Validation comparison diagrams for S1 models with different parameters. (a)--(c) Validation comparison diagrams for S1 model using 1-h data collected in May, 2012 at Lijiang station; (d)--(f) validation comparison diagrams for S1 model using 10-min data collected in May, 2012 at Lijiang station; (g)--(i) validation comparison diagrams for S1 model using 1-h data collected in July, 2012 at Kazuo station; (j)--(l) validation comparison diagrams for S1 model using 10-min data collected in December,
| Model | Training data | a1 | a2 | a3 | a4 | a5 |
|---|
| S0 | 1-h data | 0.2258 (0.2057, 0.2458)(with 95% confidence bounds) | -0.7401 (-0.7886, -0.6916)(with 95% confidence bounds) | -5.141 (-5.955, -4.327)(with 95% confidence bounds) | 8.406 (7.116, 9.696)(with 95% confidence bounds) | | | 10-min data | 0.1949(0.1698, 0.2201)(with 95% confidence bounds) | -0.8155(-0.8903, -0.7407)(with 95% confidence bounds) | -3.121(-3.639, -2.604)(with 95% confidence bounds) | 5.446(4.609, 6.283)(with 95% confidence bounds) | | | 1-min data | 0.2146(0.2089, 0.2203)(with 95% confidence bounds) | -0.7548(-0.7749, -0.7346)(with 95% confidence bounds) | -3.604(-3.803, -3.405)(with 95% confidence bounds) | 6.202(5.889, 6.515)(with 95% confidence bounds) | | | S1 | 1-h data | 0.2338 | -0.7386 | -5.5787 | 8.6573 | 0.2926 | | 10-min data | 0.2074 | -0.8187 | -3.2986 | 5.5718 | 0.1356 | | 1-min data | 0.2251 | -0.7648 | -3.9509 | 6.4351 | 0.2379 |
|
Table 1. Parameters of diffuse fraction sigmoid models S0 and S1
| Model | Training data | Validating data | r | EMBD | ERMSD | t | Skewness | Kurtosis |
|---|
| S0 | 1-h data | 1-h data collected in May, 2012 from Lijiang station | 0.9298 | 0.0406 | 0.1279 | 5.8873 | 0.9493 | 3.7014 | | 10-min data | 0.9396 | 0.0059 | 0.1147 | 0.9097 | 0.4265 | 2.9275 | | 1-min data | 0.9386 | 0.0930 | 0.1168 | 1.4137 | 0.4543 | 2.8173 | | S1 | 1-h data | 0.9325 | 0.0417 | 0.1261 | 6.1695 | 0.7483 | 3.6499 | | 10-min data | 0.9426 | 0.0105 | 0.1120 | 1.6507 | 0.2871 | 2.9649 | | 1-min data | 0.9426 | 0.0141 | 0.1128 | 2.2154 | 0.3039 | 3.0297 | | S0 | 1-h data | 10-min data collected in May, 2012 from Lijiang station | 0.9347 | 0.0299 | 0.1221 | 10.9822 | 0.8327 | 3.7552 | | 10-min data | 0.9441 | -0.0045 | 0.1124 | 1.7282 | 0.2793 | 2.8329 | | 1-min data | 0.9434 | -7.9555×10-4 | 0.1141 | 0.3028 | 0.3438 | 2.7508 | | S1 | 1-h data | 0.9364 | 0.0310 | 0.1209 | 11.5256 | 0.5323 | 3.6769 | | 10-min data | 0.9450 | 2.3221×10-4 | 0.1108 | 0.0910 | 0.0815 | 2.9916 | | 1-min data | 0.9447 | 0.0043 | 0.1115 | 1.6576 | 0.0801 | 3.1446 | | Model | Training data | Validating data | r | EMBD | ERMSD | t | Skewness | Kurtosis | | S0 | 1-h data | 1-h data collected in July, 2012 from Kazuo station | 0.9088 | -0.0128 | 0.1260 | 1.7631 | 0.2680 | 3.0993 | | 10-min data | 0.9045 | -0.0690 | 0.1492 | 8.9895 | 0.2047 | 2.4575 | | 1-min data | 0.9065 | -0.0630 | 0.1549 | 8.2535 | 0.2440 | 2.4820 | | S1 | 1-h data | 0.9255 | -0.0389 | 0.1210 | 5.8554 | -0.0042 | 3.0082 | | 10-min data | 0.9115 | -0.0833 | 0.1533 | 11.1526 | 0.1538 | 2.4256 | | 1-min data | 0.9136 | -0.0864 | 0.1537 | 11.7033 | 0.1515 | 2.4872 | | S0 | 1-h data | 10-min data collected in December, 2012 from Kazuo station | 0.9624 | 0.0128 | 0.0838 | 5.6638 | -0.3398 | 3.4018 | | 10-min data | 0.9599 | -0.0397 | 0.1054 | 14.8894 | -06306 | 2.4770 | | 1-min data | 0.9623 | -0.0333 | 0.1013 | 12.7599 | -0.5861 | 2.4319 | | S1 | 1-h data | 0.9560 | -0.0680 | 0.1113 | 28.2822 | -1.1048 | 4.6325 | | 10-1-min data | 0.9555 | -0.0929 | 0.1379 | 33.3518 | -0.6255 | 2.6717 | | 1-min data | 0.9424 | -0.0989 | 0.1421 | 35.5077 | -0.7060 | 3.0019 | | S0 | 1-h data | 1-h data collected in July, 2012 from Guangzhou station | 0.8768 | -0.0290 | 0.1263 | 4.3277 | -0.2584 | 4.2410 | | 10-min data | 0.8326 | -0.0785 | 0.1624 | 10.1203 | -0.1641 | 3.1330 | | 1-min data | 0.8446 | -0.0785 | 0.1565 | 10.6224 | -0.1482 | 3.3136 | | S1 | 1-h data | 0.8795 | -0.0240 | 0.1240 | 3.6184 | -0.2628 | 4.0508 | | 10-min data | 0.8326 | -0.0698 | 0.1590 | 8.9599 | -0.1949 | 3.1146 | | 1-min data | 0.8455 | -0.0662 | 0.1514 | 8.9060 | -0.2141 | 3.2803 | | S0 | 1-h data | 10-min data collected in December, 2012 from Guangzhou station | 0.9419 | 0.0141 | 0.0999 | 5.3019 | -0.2218 | 3.2726 | | 10-min data | 0.9391 | -0.0407 | 0.1177 | 13.6514 | -0.4412 | 2.3907 | | 1-min data | 0.9420 | -0.0342 | 0.1138 | 11.6802 | -0.3992 | 2.4080 | | S1 | 1-h data | 0.9448 | -0.0737 | 0.1209 | 28.4774 | -1.0439 | 4.4547 | | 10-1-min data | 0.9400 | -0.0964 | 0.1470 | 32.2035 | -0.5575 | 2.6305 | | 1-min data | 0.9283 | -0.1167 | 0.1658 | 36.7535 | -0.6546 | 2.9579 | | S0 | 1-h data | 1-h data collected in July, 2012 from Wuhan station | 0.9302 | -0.0480 | 0.1203 | 7.3646 | -0.2067 | 4.6677 | | 10-min data | 0.9111 | -0.0972 | 0.1569 | 13.3550 | -0.1598 | 3.2799 | | 1-min data | 0.9166 | -0.0943 | 0.1517 | 13.4203 | -0.0818 | 3.5798 | | S1 | 1-h data | 0.9304 | -0.0455 | 0.1194 | 6.9757 | -0.4988 | 4.7733 | | 10-min data | 0.9097 | -0.0898 | 0.1534 | 12.1976 | -0.3180 | 3.3477 | | 1-min data | 0.9146 | -0.0848 | 0.1475 | 11.8862 | -0.3789 | 3.7344 | | S0 | 1-h data | 10-min data collected in December, 2012 from Wuhan station | 0.9296 | 0.0335 | 0.1196 | 10.7714 | 1.6448 | 8.0956 | | 10-min data | 0.9165 | -5.8784×10-5 | 0.1257 | 0.0172 | 1.1780 | 7.8858 | | 1-min data | 0.9217 | 0.0029 | 0.1230 | 0.8583 | 1.0676 | 7.1379 | | S1 | 1-h data | 0.9177 | -0.0203 | 0.1255 | 6.0476 | 1.4370 | 9.5559 | | 10-min data | 0.9029 | -0.0354 | 0.1387 | 9.7338 | 1.1877 | 8.2914 | | 1-min data | 0.8941 | -0.0473 | 0.1475 | 12.4817 | 0.8640 | 7.1872 | | S0 | 1-h data | 1-h data collected in July, 2012 from Urumqi station | 0.9483 | 0.0091 | 0.0999 | 3.1903 | -0.5119 | 2.9069 | | 10-min data | 0.9438 | -0.0452 | 0.1280 | 13.1603 | -0.6870 | 2.2919 | | 1-min data | 0.9467 | -0.0382 | 0.1230 | 11.3951 | -0.6702 | 2.2984 | | S1 | 1-h data | 0.9479 | -0.0823 | 0.1294 | 28.8043 | -1.0369 | 4.0392 | | 10-min data | 0.9453 | -0.1022 | 0.1583 | 29.5406 | -0.6745 | 2.4640 | | 1-min data | 0.9316 | -0.1218 | 0.1761 | 33.4283 | -0.6958 | 2.7061 | | S0 | 1-h data | 10-1-min data collected in December, 2012 from Urumqi station | 0.8482 | -0.0832 | 0.1834 | 5.6485 | -0.0880 | 2.4156 | | 10-min data | 0.8258 | -0.1376 | 0.2230 | 8.6948 | -0.0104 | 2.0106 | | 1-min data | 0.8318 | -0.1292 | 0.2160 | 8.2784 | -0.0151 | 2.0626 | | S1 | 1-h data | 0.8569 | -0.0784 | 0.1773 | 5.4667 | -0.1258 | 2.2557 | | 10-min data | 0.8312 | -0.1315 | 0.2174 | 8.4246 | -0.0445 | 1.9770 | | 1-min data | 0.8394 | -0.1216 | 0.2084 | 7.9730 | -0.0731 | 2.0014 |
|
Table 2. Statistical results of diffuse fraction calculated by S0 and S1 models
