Nov. 8

Processing the flux data. It takes two steps. First the LI-7500 has this periodic spiking problem, so we are reprocessing the raw data, de-spiking it and creating new files.

1. Put raw data on D:\vaira\raw subdirectory
2. Create ascii file, FILEday.RAW, with names:

d:\vaira\raw\g2961102.RAW

d:\vaira\raw\g2961130.RAW

d:\vaira\raw\g2961200.RAW ……

Each time we down load data we will create a new *.raw file. For example the file FILE296.raw has file names for data from days 296 to 309. Next time we download data we will create a new file with the name FILE310 that will have the intervening data. This will prevent us from reprocessing the raw data over and over again, a time consuming exerecise.

3. Apply processing program

Raw_flux_grass.c despikes the data, computes 30 min mean averages, variances and covariances and output the data files similar to the GILLSONIC program.

Eg VRyr_day.flx_dspk; VRyr_day.err_dspk and VRyr_day.nrt_dspk

The output files are VR_yrday.dat by day and a total file called

Grass2000.dat with all of the data.

Preliminary examination of the radiation sensors and energy balance closure was performed.

Figure 2 Comparison of radiation sensors. The slope of Rn vs Rg seems low compared to forests (0.85) but reasonable for crops and grassland. Lots of old dry grass is showing with high reflectance. Not sure about the role of the tower on reflection. Need toassses this better.

We are concerned about the relation between PAR and Rg. Typically PAR (Wm-2) is 0.5 times Rg (W m-2). But Par (umol m-2 s-1) is 4.6 times Rg/2. The sensor is supposed to be in micromoles m-2 s-1. But its magnidude suggests W m-2. But the PAR sensor agrees with the value of the line sensor. Need to check this out more.

Figure 3 Regression of PAR vs Rg. The units of PAR seem queer

Energy balance seems excellent

Figure 4 Energy Balance Closure from first study period

The raw data files from Days 296 to 309 are in:

File296.raw

Figure 5 Diurnal pattern of CO2 exchange, initial