As far as I know ASTER might be your best bet with 90 m resolution. Surface Kinetic Temperature AST_08 products can be ordered for free. ASTER has been up there since 2000. But you need to check whether there are any ASTER scenes available for your study area. Other than that you can take a look at Landsat 7 ETM+ and 5 TM with 60 m and 120 m resolution respectively. You need to derive LST using a suitable algorithm for Landsat such as mono window algorithm. Check attached links. Where is your study area? What is the spatial resolution that you require? Are you looking for freely available or commercially available thermal satellite imageries?

https://lpdaac.usgs.gov/dataset_discovery/aster/aster_products_table/ast_08

https://www.researchgate.net/post/What_are_the_steps_to_use_thermal_infrared_band_eg_Landsat

Dear James, 

Thanks a lot for your detailed reply and attachments.

my case study is  city Beijing, China. i need higher resolution images that are freely available.  

As i checked there is no image of Landsat 8 before 2013. So i may check landsat 7.

I know how to apply algorithm for Landsat 8  image to get LST. is the procedure and algorithm same for landsat 7? ?? 

Best wishes,

Majid when you say higher resolution do you have any particular range in mind? Did you find any ASTER images for Beijing?

All these are freely available 60m (ETM+), 90m (ASTER), 120m (L5 TM), 1 km (MODIS) and there are some others. If you use Barsi et al. (2003) to derive LST the same radiative transfer equation can be used for TM and ETM+ . How are you deriving emissivity? Did you check the RG discussion link that I attached with the previous answer? 

James, resolution of 30*30 is perfect but resolution llesser than 100 m is also acceptable. I have not tried for any satellite for the years 2000,2005,2010. but i already calculated LST by Landsat 8 images for 2015 (single window method) 

its night here and i have to leave the lab now. i will check all tomorrow. many thanks for your guides. 

You may use Landsat 5, 7, 8 and ASTER imagery. Please go through these articles

http://www.mdpi.com/2072-4292/8/12/993

http://www.mdpi.com/2072-4292/8/5/413

Hello,

You can use LANDSAT TM and ETM+ for LST estimation.

Best Wishes,

Rajesh

Dear all,

thank you for your helpful replies.

Regards,

Majid

Using Aster giving you better thermal resolution 90mt. But landsat ETM, TM-4,5,, 8(60, 100) is very good resolution for change study. But choose your satellite data according to you study area. Because if study area is large or you want to study for regional scale then for you MODIS is good. So before starting your project understanding of data is very important. 

Dear Surya, 

Thank you for your answer. 

i want to estimate the LST of an urban district (about 10km*10km) in summer time. i want to compare the differences in LST in 4 years (2000,2005,2010 and 2015). i will choose one image for one year ( i only will compare 4 images).

according to my understanding from your answer, i should choose landsat 8 and  ETM and TM. right?

is it right if use and compare images of  different sensors ? i mean Landsat 8 (for 2015) and ETM (for 2005 and 2010) and TM (for 2000)? or i must use only one sensor (ETM only)?

Best regards,

Majid

Majid,

According to my opinion, it would be easier to compare images from the same sensor because of common spatial, temporal, radiometric and spectral resolution. Since the sensor specifications remain the same, you can focus on the land surface temperature sources in all four time periods. But if you deal with different sensors you will have to consider differences arising from sensor specifications in addition to differences in land surface properties. My views are open for discussion and criticism.

Like I mentioned in my first comment, ASTER might be well suited for your particular case study since the images are available from 2000 onwards. But I'm not sure if you searched for ASTER images for Beijing? It is not uncommon for researchers to work with satellite images from multiple sensors such and TM, ETM+ and OLI/TIRS, when their study covers a larger time period and for other reasons as well. However you should keep in mind that L8 TIRS has had stray light issues since it's launch which has affected both the thermal bands that ultimately biases the LST products derived from it. So when you are analysing LST derived from Landsat 8, how are you going to figure out whether the surface temperature is really caused due to the land surface properties or whether it is due to the sensor error or both?  

Dear James Varghese,

Thank you so much for your time and detailed reply.

To be honest my knowledge in remote sensing is poor. i am somehow a beginner. 

I didnt search ASTER images as Landsat providing higher resolution than ASTER.

i have no idea about stray light issues in Landsat 8 (i only heard about it for Landsat 7). i should read more apparently. 

if would be thankful if i could get more information about my question from  your and other scholars comments.

Warm Regards,

Majid

Thanks Majid for your clarification. I am also a student like you, therefore I am learning too. But whatever I know I will be glad to share.

It would be useful for you to go through the links that I have attached. The first link explains what stray light issues are in the context of Landsat 8 and what are the consequences if this issue is ignored. For example, USGS link states, 'Differences between the ground-based results and TIRS results ranged up 5 K in the TIRS Band 10 and up to 10 K in TIRS Band 11'. Some of this variation is related to the time of year (i.e., the temperature differences were larger during the summer,'. Landsat 7 ETM+ has stripping issues, haven't heard about stray light issues in L7. It is true that L7 ETM+ (60m) has better spatial resolution than ASTER (90m), however there are drastic differences when you consider spectral resolutions. ASTER has 5 thermal infrared bands whereas ETM+ only has one thermal band. This property becomes crucial when you want to separate Temperature and Emissivity which are critical for LST studies. 

I did a quick search over Beijing for ASTER images and found 358 images since 1999 until 2016. Check the NASA link that I have attached. 

If I understand correctly, you are using one LST image each for the year 2000, 2005, 2010 and 2015, right? What intrigues me is, how can the Land Surface Temperature taken over an area at an instantaneous moment in time be representative for the entire season, or year? Is it okay to conclude using one LST image for a year that there is a rise or fall in the LST trend? Don't you need to consider more LST images to establish a trend over a year or for years? I am just curious. I don't know the answer myself but I just had these questions in my mind. Apologies if I have more questions for you than answers. What does your reading of other research/review of articles on Urban Heat Island Effect and Cooling Effect of Green spaces tell you? 

https://landsat.usgs.gov/l8handbook_appendixa.php

https://search.earthdata.nasa.gov/search/project?p=!C14758250-LPDAAC_ECS&m=39.9074977414405!114.72291183471697!7!1!0!0%2C2&tl=1007720730!6!!&q=ASTER&ok=ASTER+over+beijing&sb=115.81419234604371%2C39.463175993293206%2C116.98026413589028%2C40.3518194895878