Land surface temperature modelling using surface energy balance is possible due to calculation of surface temperature from the sensible heat flux: H = ρcp(Ts - Ta)/ra. It seems to be simple, however it is very complicated because you need know ra (aerodynamic resistance of the surface for the energy and momentum transfer) and H. In context of Monin-Obukhov similarity theory, ra depends on temperature gradient (i.e. surface temperature). I think, it is simple to measure surface temperature, modelling of surface temperature is crazy:-)

I think you should start with data that you have and from which you want to estimate surface temperature. Sometimes it could be possible to estimate this parameter without using energy balance equation. For instance, when you have measured wind and/or air temperature at more than 1 levels you can use Monin-Obukhov similarity theory (MOST) to estimate heat flux with the so-called 'profile' method. Then you can estimate soil temperature. However if you have 'minimum' set of observations, namely  wind, temperature, humidity (dew point) each measured at single level and cloud cover you have to use the full energy balance approach with several involved parameterizations. I would not say that this is completely meaningless excersize - sometimes it is have to be accomplished for instance to initialize atmospheric dispersion models, etc. Review of this subject is available in the Final Report of COST  710 project "Harmonisation in the pre-processing of meteorological data for dispersion models"  available here http://www2.dmu.dk/atmosphericenvironment/cost_710.htm#Results

Thank you Dr. Jakub, Dr. Sasi and Dr. Ivan for your comments.

Let me add some more 'craziness' to the equation. The ultimate aim of the research problem which I am tackling is to see how 'soil moisture affects geothermal surface temperature anomaly detection via remote sensing techniques'.

Geothermal surface temperature anomalies are thermal anomalies associated with geothermal features which have sub-surface heat source. Identifying these anomalies are challenging using thermal infrared (TIR) remote sensing techniques because 1. these anomalies are relatively small compared to the size of a pixel (90m by 90 m) and 2. they do not show high temperature contrast with the surrounding region. Detecting them via satellite remote sensing is complicated also because the external source of heat from the sun has to be separated from the surface temperature associated with the geothermal source. Now this is where the picture becomes even more complex. When soil moisture effects are added, the hypothesis is that it obscures the effects of thermal emissions from these geothermal features. So to separate the temperature effects of the sun and the soil moisture temperature effects from the geothermal temperature effects the most likely path which I could think of is to model the three aspects. Ideally I would like to quantify the temperature effects caused due to soil moisture and separate it from the geothermal source. Moreover, I do not have any ground validation options. Therefore modeling seems to be the best option I have with the support from literature. It has potential applications in geothermal exploration and volcano-geothermal hazard monitoring..

I would like to understand how soil moisture affects land surface temperatures. If you can shed some light here would be helpful.

Dear James,

it is a bit tricky to make general and simple quantitaive statements about influence of soil moisture on soil temperature even though there are several widely known qualitative facts that amplitude of diurnal temperature variations increases when soil moisture decreases, etc.

There are several old theories of diurnal temperature cycle under idealized conditions in which (in particular) effects of soil moisture on temperature are quantified. Such theories are useful for physical understanding, but real atmosphere is often being far from the involved assumptions (e.g. small temperature change due to synoptic forcing)

I think that for your purpose you could use some well established land-surface model which can be driven by standard measurements of synoptic meteostation. One of such models which freely available is NOAH LSM http://www.ral.ucar.edu/research/land/technology/lsm.php

Dr. Ivan , I also read some papers on the effects of soil moisture on land surface temperature. They say that soil moisture decreases the diurnal LST range. Logically this sounds to be right. I am also making the same hypothesis for my research problem but I will have to test this proposition in active geothermal areas. But you are right, the truth is far from the assumptions we make. What is a synoptic forcing?

A brief look at the link which you shared seems quite helpful. I will look at it in detail and try to see if it can be applied in this research problem. Thanks for sharing the link.

Existing theories of diurnal temperature cycle for which analytical solutions were obtained consider air and soil temperature changes due to diurnal changes in incoming shortwave radiation. They use energy balance equation together with the equations of heat transport in soil and in air . However temperature could change not only due to radiation, but also due to movement of air masses and passages of cold and warm fronts. This is what I mentioned as 'synoptic forcing'. Therefore practical usage of such theories is limited.

Dear James, your task is very complex and also crazy. On the other hand, crazy things can have the best results:-) I thing energy balance approach is too complex for modelling surface temperature considering lot of difficult inputs and relationships. For example surface resistance for heat transfer depends on temperature gradient in the atmosphere, surface geometry, on the character of turbulent flow of the air in the atmosphere, i.e. atmospheric stability, etc. Errors of the calculated results of the energy fluxes from the measured data are usually in range of 10 to 20%. Furthermore, important impact on energy balance can have vegetation which can change relations (ratio) between energy fluxes. Vegetation cover can also mask impact of geothermal surplus energy due to transpiration (physiological cooling of plants surface). On the other hand, vegetation cover can indicates areas with higher (lower) soil temperature - vegetation can have higher biomass, different species composition and different spectral features (usable in RS). One of the possibility for analysis of the relations between soil moisture and thermal properties of the soils (ground) is usage of thermal inertia approach - see papers of Antonino Maltese from Palermo University.

Good luck!

Jakub

Thanks Dr. Jakub, I have been looking at various methods and one of them is Apparent Thermal Inertia (ATI) = 1 - Albedo / dT

Still trying to figure out which one will be most suitable.

Dear James, ground surface temperature can be obtained from coupled models of for energy balance, atmospheric turbulence, soil water and heat dynamics, and plant root water uptake (e.g., https://www.researchgate.net/publication/236017872_Development_of_a_soil-plant-atmosphere_continuum_model_%28HDS-SPAC%29_based_on_hybrid_dual-source_approach_and_its_verification_in_wheat_field ). For bare soil, the model will be much simpler.

Article Development of a soil-plant-atmosphere continuum model (HDS-...

Thank you Prof. Shang for the comments and for sharing the article.

usually, we go the other way. We estimate surface temperature from the thermal band and estimate latent heat or sensible heat using the surface energy balance equation. If you have eddy covariance data that measures or estimate the energy balance components, then you can estimate surface temperature from the sensible heat equation (delta T)

Thank you Prof. Melesse for your answer. Makes perfect sense.

dear James

 please follow the link 

https://www.researchgate.net/publication/301797360_Analyzing_land_surface_temperature_distribution_in_response_to_land_useland_cover_change_using_split_window_algorithm_and_spectral_radiance_model_in_Sundarban_Biosphere_Reserve_India

Article Analyzing land surface temperature distribution in response ...