I’m planning to investigate the variability o f the mixed layer depth in the Moroccan Atlantic cost using CTD data. I would like some advice about recommended techniques. I’m really looking to find out what you’d consider to be the most efficient method.
Hello Ismail,
with your CTD data you can use different threshold criteria in order to find the mixed layer depth. Personally I prefer the density threshold criterion of 0.125 kg/m^3 for the North Aegean Sea. So it is quite straightforward actually:
You first find the potential density anomaly (calculated based on temperature and salinity e.g. by using the TEOS-10 GSW sea water toolbox) at the surface and then you add 0.125 kg/m^3 to this values.The mixed layer depth is the depth (based on your pressure values) where the density anomaly is equal to the density at the surface plus the density threshold:
MLD = depth where(σ0 = σ0(surface)+ 0.125 kg/m^3)
What software are you using?
There are some scripts available for Matlab
https://github.com/rsignell-usgs/mld_demo
I hope that helps,
Naomi
Dear Naomi Krauzig,
Thank you for your answer.
Actually, in a work that will be published soon, we compared the MLD using the threshold method and the gradient method of temperature and density from CTD profiles in-situ data.
To find the approximated MLD of the temperature and potential density profiles starting from the surface, the threshold method searches progressively deeper levels until they find a level where the temperature or potential density differs from the surface reference value by a specified threshold. In our study, we used (0.3°C) for temperature and 0.03 kg.m-3 for density as the threshold difference criteria and we used 10 m as the surface reference level.
The Gradient methods, which also frequently used, work much like threshold value methods. It assumed that there is a strong gradient at the base of the mixed layer and therefore search for critical gradient values. The commonly used values ranges from 0.0005 to 0.05 kg.m-4 for density gradients and 0.025 °C m-1 for temperature gradient. we used 0.005°C.m-1 and 0.0005kg.m-4 as the temperature gradient and potential density gradient criterions successively.
basically we found that the threshold method of temperature (0.3°C) is the best method for studing the variability of the MLD in the Moroccan Atlantic coast.
I am using Maltab for my studies.
Remark: The water masses in the Atlantic coast of Morocco is so different from the water masses in the basins of the Mediterranean Sea. for example, the vertical distribution of salinity is reversed in the Atlantic Ocean relative to salinity in the Mediterranean. in other words, the highest salinity is at the surface and the low salinity at the depth in the Moroccan Atlantic coast.
This is my email if you wanna discuss more about the variability of the MLD or for a future collaboration : [email protected]
Ismail
Ismail Bessa Naomi Krauzig hello, I have the same question. I want to get MLD from ctd profil from the Mediterranean Sea and I am not sure how to choose the best criterion since they are giving different results for MLD. Can you please explain why you prefer 0.125 for the Med sea Naomi Krauzig ? Thank you
Nawal
Hi Nawal Bouchachi, In my studies on the MLD in the Moroccan Atlantic coast I am using 0.5 °C as a threshold value for temperature with 10 m as the surface level reference.
I suggest, to calculate the MLD in two steps. First, assess the MLD visually, based on the CTD profiles(T, D). second, apply the different threshold values to find the one that is adequate to the MLD that you visuallazed.
Cordially
Nawal Bouchachi I am sorry I should have been more specific.
I meant the North Aegean region. Based on bibliography maybe you can find the criterion for your region of interest?
Here is on article for the Mediterranean:
Article Seasonal variability of the mixed layer depth in the Mediter...
As Ismail Bessa mentioned the 10m reference is important.
Here are also some indications from Ifremer:
MLD is defined through the threshold method with a finite difference criterion from a near-surface reference value. A linear interpolation between levels is then used to estimate the exact depth at wich the difference criterion is reached. The reference depth is set at 10 m to avoid a large part of the strong diurnal cycle in the top few meters of the ocean. The chosen temperature criterion is 0.2°C absolute difference from surface (DT). The fixed criterion in density is 0.03 kg/m3 difference from surface (DR with R for Rho i.e. density). The variable criterion in density corresponds to a 0.2°C variation in local temperature/salinity conditions, it is obviously linked to the temperature criterion and enables the definition of the BLT. (See de Boyer Montégut et al. JGR 2004, for further details about the choice of the criterion).
http://www.ifremer.fr/cerweb/deboyer/mld/Surface_Mixed_Layer_Depth.php
I am using 10 m as reference depth as well. So, thank you so much for your indications for criterion value choice!!
Nawal
You may be interested in Holte and Talley 2009 paper mentioned here http://mixedlayer.ucsd.edu/
The algorithm is part of the Climate Data Toolbox. It is based on both the threshold and gradient methods for T, \rho and S, and chooses the "best" one. If only T data are available, uses that instead but choosing the best of the gradient vs threshold values. It seems worth it cause neither criterion are solid always.
https://www.mathworks.com/matlabcentral/mlc-downloads/downloads/dff4ffb7-190a-4494-839f-ff39787bd48e/eb22d828-c61c-430d-b325-e6e020330325/previews/cdt/doc/html/CDT_Getting_Started.html#2
I just found this and plan to use it.
Cheers
Paula,
Clément de Boyer et al., (2010) "Mixed layer depth over the global ocean:
An examination of profile data and a profile-based climatology"
Also can be of interest !
Abrazos,
That is a very nice toolbox indeed, but depending the region and season you might want to nudge the threshold. I find that 0.03 does not do it for all oceans :)
Saludos!
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA