Adjust the BUMP NICAS Vertical Localization Scale for HAFS Test Case

[JingCheng-NOAA]

The previous settings of vertical localization scale in BUMP NICAS yaml file tends out to be too big for vertical localization, as mentioned in #6 and #13 .
This issue records the process of finding the best number for current HAFS test case (81 vertical levels).

[JingCheng-NOAA]

For now, only default sigma level option works in JEDI for FV3 model. Previous experiments were using 0.3, which turned out to be too broad vertically in terms of increment brought in by single observation.

[JingCheng-NOAA]

Testing scale number for Sonde data (temperature on 850 hPa)

[JingCheng-NOAA]

Testing on surface temperature data.

[JingCheng-NOAA]

To reduce the complexity in test cases, the number 0.01 is chosen to be as the number for vertical localization with HAFS model that use 81 vertical layers.

[JingCheng-NOAA]

Though single obs test prove the number 0.01 is enough, but multiple sondes data (ADUDPA airTemperature) test show the increment on vertical levels are limited to lower level. Need to revisit this issue to check out why.

[JingCheng-NOAA]

Adjust vertical localization scale from 0.01 to 0.1 will increase the levels of impact, but also increase the magnitude of increment.