Overlap correction algorithm

Algorithms for inferring overlap correction for (CHM15k) ceilometers from atmospheric observations under well-mixed boundary layer conditions. Based on Hervo et al. 2016 (https://doi.org/10.5194/amt-9-2947-2016)

Content

• Calculate overlap correction funtion ( script_overlap_routine_v2.m calling calculate_overlap_automatic_structured_2.m)
• plot all results (read_overlap_cor_v5). Including:

1. All daily overlap functions
2. temperature model
3. Impact on gradient and temperature model simple PBL detection

• Gradient Analysis (read_gradient)

Detailed description of standard way of using

1. script_overlap_routine_v3_EPROF.m
	a. fill variables : 
		* stn: wigos-id
		* start_time/end_time: time period for which to calculate overlap correction
		* folder_data : location of E-PROFILE netCDF level 1 files   (NOTE: you'll prefer to transfer ncdf files directly on your machine to improve execution speed (locally ~12hours/1year of data))
		* folder_out :  output directory   (NOTE: prefer a local directory on your machine for speed)
	b. run script                              
2. read_overlap_cor_v6_eprofile.m
	a. fill variables:
		* station: wigos-id
		* folder_ncdata:  location of E-PROFILE netCDF level 1 files
		* folder_correction:  File path of the output created by script_overlap_routine_v3_EPROF.m in step 2)
		* folder_output: output directory  (NOTE: prefer a local directory for speed)
	b. If you are configuring a new station add a case in the station inputs switch. Then specify:
		* timerange of the correction files (in info.xxx) + timerange of daily correction visualization for actually testing the correction on measured profiles (in info_test.xxx).
		* optical module id (info.tub) and instrument serial number (info.chm). You'll find this information in daily netcdf files attributes.
	c. run script             
	d. filter bad quality overlap estimates
		* identify in plots bad quality dates by using the overlap function vs range plot (Figure 2: fig 4) and the dates indicated in the command line output
		* add these bad days to list_dates_bad_quality in the station switch
	e. run script again

Examples

Raw


Corrected with model