The term desertification seems to mean different things to different people. Broadly defined as reduction of the biological potential of land in arid and semi-arid areas, there are three main effects: change in aridity, vegetation species change, and change in soil properties. Most remote sensing focusses on change in 'greenness', for example defined by products such as LAI, or indices such as NDVI, which relate well to GPP, and usually correlate closely with precipitation variation in semi-arid regions. Here a long time series is needed as typically there is high inter-annual variation. Species change may often be more important, for example change from grassland to woody shrub invasion. This may not show up if we look only at coarse scale, so finer resolution and classification techniques are needed. For soils this is more difficult - there may be some change in soil colour with salinity, but usually vegetation productivity is used as a surrogate for other stress factors.

@Behrooz Arastoo, color infrared remote sensing would be the best choice.

You mean desertification or aridity. The aridity can be easily observed on Landsat Aster MODIS Misr and several others essentially look to SAVI and NDVI.

Hi,

you can also consider satellite soil moisture products for your purpose,

They are freely available on a global scale with daily temporal resolution.

See the reply to this question (https://www.researchgate.net/post/What_is_the_best_satellite_soil_moisture_product_for_land_surface_modelling?) for an overview of the available products.

Kind regards,

L

Luca was this product radar derieved? certainly the soil moisture can be taken as an indicator for change in aridity over years.

by integrated approach only desertification from image will be made.... first identify the main process, type and severity ..base upon the area and objective of the study we will go for suitable image/satellite data.....

Hi Saif,

these satellite soil moisture products are obtained by passive and acrive microwave sensors, i.e. by radiometers and scatterometers.

Also evaluate the time series of vegetation indices (i.e. NDVI) &/ or NPP. Decreasing trends in vegetation conditions and thus productivity over a long term can be used as an indicator of desertification

Yes Ranjani this is what precisely I indicated, however you really need a long time series, else it can be deceptive. there are both wet and drought years on decadal scale.

The term desertification seems to mean different things to different people. Broadly defined as reduction of the biological potential of land in arid and semi-arid areas, there are three main effects: change in aridity, vegetation species change, and change in soil properties. Most remote sensing focusses on change in 'greenness', for example defined by products such as LAI, or indices such as NDVI, which relate well to GPP, and usually correlate closely with precipitation variation in semi-arid regions. Here a long time series is needed as typically there is high inter-annual variation. Species change may often be more important, for example change from grassland to woody shrub invasion. This may not show up if we look only at coarse scale, so finer resolution and classification techniques are needed. For soils this is more difficult - there may be some change in soil colour with salinity, but usually vegetation productivity is used as a surrogate for other stress factors.

@ North, I agree with some of your concerns, i.e. changes in species composition vs. scale. But, long term trends in vegetation (i.e. NPP), even at medium to coarse scale, may reflect these changes in species composition. For example, if over the periods of 1-2 decades woody encroachment in grasslands can be seen in increasing trends in NPP. Also, extensive research has been shown that annually/ seasonally integrated NDVIs correlate well with NPPs (field measured &/ remotely estimated), so long term records of NDVIs can still be a good choice

@ Behrooz, if you are interested in time series of NDVIs, you can look for AVHRR GIMMS NDVIs. They are available as bi-weekly MVCs globally at 8km spatial resolution for 25 year period from 1982 -2006. Also, MODIS derived NPP data are available globally at 1km spatial resolution from 2000 to 2011

Howdy.

Soil Moisture Content (SMC) can now be measured based on thermal inertia quite accurately. Hence follow SMC in the region of your interest and SMC will give you inormaton

Dear Frank

thank you so much

can you explain more or give me more reference and information.?

Hi Arastoo BEhrooz,

I can refer to two papers describing the approach of how to use thermal sensors to estimate Soil Mositure Content. On describes the theoretical basis, the other is a validation of the technique performed in the Xinjiang province in the Northwest of CHina witha very arid cliamte, but with some very high mounatin ranges which provide water for irrigation (of cotton) and for drinking (evidently)

Feel free to ask questions of something is not clear in these papers.

Here are the links:

https://www.researchgate.net/profile/Frank_Veroustraete/publications?pubType=article&ev=prf_pubs_art

https://www.researchgate.net/profile/Frank_Veroustraete/publications?page=3&sorting=newest&pubType=article

Cheers,

Frank

Howdy Shariar.

As soon as your sensor has thermal channels you can quantify soil moisture content (SMC). This means for example that global SMC can be produced (which has not been done yet) wit a backcasting perspective of more than 25 years. Don't try this with a radar passive or active system. The datasets don't go back that far, and show gaps between missions. The spatial resolution of SMC derived from thermal inertia is 1x1 km, that's pretty much better than passive and even active radar systems products (40x40 km).

Moreover, due to the low spatial resolution of the radar systems, they are prone to signal contamination with water surfaces on land or along coast lines. In Europe one has to throw away a huge amount of coastal data due to this pheneomenon with radar systems.

This problem is absent when thermal inertia is applied. Hence higher spatial resolution and higher quantitative accuracy is obtained. I agree that cloudiness reduces the number of usefull data as well with SMC based on thermal inertia.

Last but not least, it is only the diurnal difference in temperature which is usefull for SMC determination, NOT as you suggest LST or emissivity becasue they are physically unrelated to thermal inertia, by definition even. Also MODIS surface reflectance or its soil moisture indices are no poxys at all for a direct measurement of SMC or thermal inertia, which is directly linked to SMC.

Not convinced. That's your good right, but then first read the papers which I have published on this topic (See ResearchGate). The technique has been applied and validated in a 200.000 km² large arid province of China. It worked perfectly with locally received MODIS AQUA and TERRA data, and is based on physics not on empirical indices or optical reflectance which are physically unrelated to SMC.

Microwave is an alternative but a less optimal one for the reasons I mentioned, water surfaces give a very large bias in that type of SMC remote sensing.

Cheers,

Frank

Dear frank

thank you so much for your kindly attention

my phd thesis was detect of desertification using remote sensing

u can by

1- using landsat or modis images

2- using programm like envi or erdas

Dear Fadia W Alazawi 

thank you so much

which RS Product is suitable for detect ?NDVI ,soil moisture  or land cover.

True to the fact that there is no reference scale for Desertification. You may better use meteorological approach to derive aridity index and so on. arriving to the conclusion of the region possessing higher aridity index. Then reorient to Remote Sensing approach of deriving vegetation indices of various time scales. If you can integrate the meteorological indices with Remote Sensing indices a better results could be derived.

the modis data can meet it

Today, there are many studies as user remote datas in detection desertification problems (NDVI, climatological indexes, that user precipiation and temperature air in surface, evapotranspiration, etc). Is important knew the environment conditions to associated the climate variability of the indexes.

Detect desertification by RS is combination of visual interpretation, knowledge about area and physico-chemical properties of soil which include ground data collection.

Based on scale of mapping u can choose image...If u want 250K scale pfo mapping u go for 100m/250m data...If u want 50k scale of mapping go for 25m data (eg liss3)....For 10k map go foe 5m/1m data....

Hi Behrooz,

I made a study of desertification in the Xinjiang province (North- Western China) lately, using a  multi-temporal series of 50 years Corona and Landsat imagery. The LUCC is enormous over that time lapse in the Xinjiang province. It led to increased desertification in originally natural landscapes. In short, an originally natural landscape  with gallery forests along the Tarim river and its tributaries changed into a patchy landscape with irrigated agriculture plots on the one hand and sandy deserts on the other hand. This LUCC had a fundamntal impact on regional hydrology. The Tarim river in the 50 years mentioned retracted by more than 200 km from its original Gobi desert ending.  If you want to know more, then read the paper available on ResearchGate:

https://www.researchgate.net/publication/235959317_Diachronic_Mapping_of_LUCC_in_the_Northwest_of_China_over_the_last_half_Century_Conversion_of_CORONA_Panchromatic_reflectance_into_LANDSAT_NDVI_Imagery

I think that the Xinjiang province gives a fairly good example of what large scale desertification is all about e.g., humans changing natural landscapes - in equilibrium with regional hydrology - replaced by desert and irrigated cropland (mainly cotton for export to SoutWestern China sweatshops). In fact the original water resources in the new landscape has been focussed on irrigated agriculture and the original natural landscape  degraded into sandy desert. Another large scale example exists in Kazakhstan (Aral Sea), not that far Norrth from the Xinjiang province, with the difference that the Aral Sea has almost vanished into pure desert. Al fishing activities in the Aral Sea have stopped and people had to move or die.

Cheers,

Frank

Article Diachronic mapping of LUCC in the northwest of China over th...

see at : https://www.researchgate.net/publication/303057047_A_spatial_desertification_indicator_for_Mediterranean_arid_rangelands_a_case_study_in_Algeria 

Note: NDVI is useless in arid - desert - semi-desert -steppe environments because of soil reflectance interference ... but so easy to use at the tip of finger on computers ....! 

Regards 

Gus 

Article A spatial desertification indicator for Mediterranean arid r...