Land Use Science - Science topic

Dear Shonam Sharma,

Data may download through WMS/WMTS in ArcGIS and QGIS and you can follow this video link to get data from Bhuvan.

https://www.youtube.com/watch?v=0098p61dQuM

Cheers.

Rajasekhar

See my answer to your personal message. Also look at the excel example there and on the previous answer above. Note you have to have the SOLVER ADD-IN activated in EXCEL. The idea is to have an objective (like minimize costs) and constraints.

you can download data from

Here it the large file document

I looked at table, the data is column separated, and the daily times were so unusual.  Then as suggested in your question, there were 31 codes of columns with data in it, probably some flow or other indicator.  Since I do not know Spanish, I could not decipher details, but the formating of the daily data in 31 columns verified the daily max, min and average, I could not figure out the 31 daily codes, what they might be (possibly discharge readings at different times), and why the three readings per day, one with no time, one with 0700 hours and one with 1700 hours.  The 31 codes (coda) were possibly a measure of discharge during the day. 

I would suggest contacting the agency collecting the data for an explanation.  The format may be clear to them, but they might not be aware that its formatting is creating issues for others in its reading and interpretation.  

In downloading data from USGS website, it is not uncommon to have to adjust the date time indicator into separate columns, and select out specific data if you want only date, time, and stream discharge, there are several columns such as stream stage removed.  I usually have a spreadsheet with original data, and then I copy and paste into another spreadsheet for alterations, adjustments for analysis, plots, etc. 

First, define an optimal solution. I would say that an optimal solution is one in which the price of land is equal across land use zone interiors (ceteris paribus). Second, develop a procedure by which the optimum can be achieved. I would say that this would require information on the elasticity of demand for land in the various zones as well as the existing prices. You might read the following for an introduction to the optimality condition: https://www.researchgate.net/publication/265426821_The_Value_of_Zoning

Hi Bibi - Its good to take as much as possible points for high accuracy and the number of pixels in one point must be n(n+1), here n is number of classes.

Just a word on how to use land use transition matrix.  I cited it, firstly, as a simple key word into relevant literature.  But it has some valuable content of its own.  Land doesn't just change use--it changes from one use to another.  Urban land is very unlikely to change into forest land because of both land values and transition costs.  Cropland can go in and out from grazing with little cost from year to year.  That's one reason why Conservation Reserve Program in US is so popular with farmers. Land can go from crops to trees (either forest or plantation) but return will not be realized for 10-40 years.  It is much less likely to go back.  Naveen wisely brings up climate change.  It can affect profitability of all land uses, but the first thing to look for is whether it pushes any single use below the margin of profitability.  It is also worth looking at regions where lots of land is at one or more margins.  There is little chance that downtown London will go into grazing nor that corn/soybean land in central Illinois will become forest.  I looked at a region at the margin in my 1985 book, COMPETITION FOR LAND IN THE AMERICAN SOUTH.  It has lots of hypotheses on why land changes use, but also many thoughts on the important question of WHERE.

Maybe "inter-regional land use" (as suggested by D. Agyemang above) or "non-local land use" in contrast to "local land use". The ISO standard for land administration, the Land Administration Domain Model (LADM, ISO 19152:2012) only use land use as a general term and does not specify regional / local land use. The standard mentions local, regional and national levels, but does not provide any definition of them. They are common terms and may be used differently around the world. It is therefore difficult to define them in an international standard. Since there also not seem not to be an established terminology (native English speakers may correct me on that), I think you can choose a term and use it in your article / paper yourself, as long as you provide a definition / explanation of it in e.g. a footnote. You should also define what you mean with "region", for example an administrative/political or geographical entity, depending on the content of your publication.

Great question and suggestion above.

The Olofsson ref below is a good one on rationale and methods for accuracy of change maps. Also recall, that if you are interested in tracking change over time it is not recommended that you compare maps post-classification. A spectral change detection stage will be required; see link to Fuller paper. 

Since both variables are categorical, I recommend Spearman's or Cramer's Coeficient of Correlation.

The Cramer's Coefficient is implemented in the MOLUSCE plugin for QGIS.

Hello Aldrin Mwilima, you may follow my research articles on land use conflicts (through the links bellow)....., will help you for both qualitative and quantitative analyses.

Habib

Thanks to all for your help at your end....

Dear Francisco,

I also consider boreal forest to be important "candidate", but there are more and less vulnerable among them. The one of the most vulnerable are East-Siberian larch forests on permafrost, because although the trees are usualy survive, the groud layer of vegetation and soil are affected a lot and don't restore then. More, even after fire the permafrost layer continue to erode because the ground layer plays a great role in maintaining of temperature regime.

I guess I could provide some literature on this issue if needed, but I have it just in Russian. Please let me know in you need it and ready to read in Russian.

Best,

Both abandonment and degradation needs to be made specific for any scientific statement. 

Abandonment of centuries of sheep grazing leads after half a century to (re-?) expansion of beech forest in the montane belt of the Apennines (attached). The impact is manifold. Pasture plant species decline in number of specimen, autumn/winter food (acorns) for deer and bear increase, absence of shepherd dogs facilitate the spread of bear and wolf.

In the arctic-alpine belt , Pinus mugo, in itself rare in the Apennines (not in the Alps), spreads replacing open vegetation with rare endemics (attached). Assessing such changes as degradation depends on priorities.  

There are several methodologies which will help assessing the impact...for understanding dynamics... application of satellite imagery, if possible HRSI, use of PRA/PLA for mapping with local community involvement, data from Agriculture department, etc. Therefore, integrated approach is most  useful.

hi

yes why not

Hossam

subtract or merge (depending on interested output) muti-temporal images layers for LULC. You can do it using ArcGIS v10+

Food and Agricultural Organization (FAO) has some old estimates. Because I am not sure of any recent estimates for these countries, I would suggest that  you use the FAO data with some caution due to the dynamic nature of land use.  

Where Q is an integer representing the bits. LISS III is 8 bit data.

Lrad = (DN/MaxGray)*(Lmax – Lmin) + Lmin.                   

Lrad: Radiance for a given DN value.

DN: Digital Count.

MaxGray: IRS-1C/1D: 255 for Corrected products.

For IRS-P6: 255 for LISS-3&4, 1023 for AWiFs for corrected products

Lmin/Lmax: Minimum/Maximum radiance value for a given band and unit is mw/cm2/str/um.

  E λ = E0 [ 1.00011 + 0.034221 cos x  + 0.00128

 where,          

E λ : Irradiance value in W/m2

 E0 : Average irradiance 1366.9977 w/m2.

 x = 2π (dn - 1) / 365

                                     r λ = Lλ / Eλ                                    

where,

r λ : spectral reflectance is expressed as a percentage.

L λ  : Radiance for a given DN value.

Eλ   : Irradiance value of particular day.

                   Ρ = ∫ r λ E (λ) dλ                                        

where,

Ρ       : Bidirectional Reflectance Distribution Function.

r λ       : Spectral reflectance is expressed as a percentage.

E0 (λ) : Solar irradiance of each wavelengh band in w/m2.

                      α(λ,θ0) = Fu(λ,θ0) / Fd(λ,θ0)                              

 where,

 α(λ,θ0)     : albedo value in no unit.

 Fu(λ,θ0)   : upwelling radiative flux in mw/cm2.

 Fd(λ,θ0)   : downwelling radiative flux in mw/cm2.

 λ             : wavelength in micrometer.

 θ0                    : solar incidence angle in degree.

U try using above formula in Envi band math or Matlab

Dear Vikram

1- Image acquisition

To conduct a LCLU change analysis, you can employ  landsat data (Free) having 30 m reslolution or Spot data (paid) with 10 m resolution even you can acquire LISS4 or LISS 3 data also if you have sufficient funds. 

2- Image Classification

Simple maximum likelihood/ complex random forest or Support vector machinealgorithms could be used to classify imageries of different years and later on comparative study could be performed to quantify changes.

3-To study impact of LULC on runoff 

The SCS-CN methodology is widely used. You need  (a)soil hydrological groups

(b)land use categories (c)tables with CN (curve number) values (d) measured rainfall.

Then define Antecedent Moisture Condition (AMC) 5 days prior to designed rainfall or measured rainfall and modify CN values according to AMC I, II or III

- use e.g. ArcGIS and HEC-GeoHMS extension to create CN-grid

- them you can create rasters of potential retention, surface runoff depth, contributing areas, surface runoff volume, etc. according to the formulas presented e.g. in the cited works

- compare the results for different time horizons with relevant land use

- you can quantify changes using map algebra in ArcGIS.

(Source : Matej Vojtek, Univerzita Konstant’na Filozofa v Nitre).

Following links may be of use.............

Actually similar problem was mine also. see the following link

How to quantify the impact of land use change on surface runoff or flooding? - ResearchGate. Available from: https://www.researchgate.net/post/How_to_quantify_the_impact_of_land_use_change_on_surface_runoff_or_flooding

Best wishes

The relationships between HD, water use & LU are direct, many & varied. The threats on water increase with time, as the stresses of impacts of climate change under a changing LU pattern accompanying changes in HD are occuring. I will send you attached a paper we presented at WATERMED Conference in Morocco in 2006 that would shed some light on your question. It also focuses on using remote sensing for the study.

 you could see The International Centre for Integrated Mountain Development web site:

There is a nice chapter for comparing the classification techniques accuracy in fallowing link:

Grass has one (sort of) G7:v.net.salesman:

A couple are listed at spatial analysis online.

I'm going to follow this as well - my current research will focus on the expansion of WUI communities...

Dear Maher,

It's difficult to help you because you do not provide enough information on your problem. If you want to understand that equation specifically, could you please indicate what S1, S2 and n are supposed to be? Also, the equation as written is potentially ambiguous or may have extraneous terms.

If, on the other hand, you are interested in land degradation and have data for an area taken at two different times, then describe what evidence you have, what you are trying to achieve and what sort of help you require.

Cheers, Michel.

You may contact National Bureau of Soil Survey and Land Use Planning at Nagpur, or its  regional centre at Delhi. Google it.

I concur with Nathan that Marxan or a similar program could help with this. I would also add that once the data is prepared and an objective function is defined, one could use optimization in any program (such as R). The key would be to get the spatial information together in the required formats, which is what Marxan and other programs provide assistance with. 

There are some wonderful new tools for doing this.  But just a caution not to get so caught up in the data that you forget the theory you want to apply.  For example, you might identify several great habitat areas in the region you are analyzing, but you need to grapple with the classic "one large vs. several small areas" issue that is fundamental in the conservation biology literature.  Also, questions of corridors, edge vs. interior habitat (the latter a function of area size), crops or trees that act as barriers to movement of target species vs. ones that are permeable.  Even if you are only looking at agriculture or forestry, you have to consider minimum sizes of fields/plantation areas and length and location of haul roads.  For some applications, you might want to go back to the 50 year old ur-source for suitability analysis, Ian McHarg's Design with Nature for a general framework.  The idea of overlays and assigning suitability points to given areas has much to commend it.  I'm encountering some of these issues in trying to analyze the suitability for development of a 40 acre suburban parcel where more than 200 townhouses have been proposed. The area contains a stream floodway (and floodway fringe) and large trees.  Among the constraints are locating a long distance trail through the property and the demands of neighbors on one side for as large a vegetated buffer as possible.  Computer mapping can be helpful, but it doesn't give a single correct answer.  One has to run through various scenarios, with some constraints absolute (no building in the floodway) and others involving tradeoffs (saving big trees away from the neighbors vs. buffering the neighbors from the new building).

Please go through the attached links they may be useful to you.

Malaysia has implemented different fuel prices for different sectors, and it has lead to the blackmarket sales of lower priced diesel to other sectors.  This was done with nano-tagging and investment in special equipment to identify the naon-tagging and after huge investments, the system eventually was dropped.

Now they are planning a new multi tier fuel pricing based on individuals income.  Which means, each time I go to fill petrol or diesel, I need to show my identification card.  That is somehow linked to my income, and that determines the fuel price I will have to pay for my purchase of fuel.  It has not been implemented yet, but this is what the rumour mill is saying.  Of course, the very rich, will have drivers to fill the fuel into their vehicles.  And if the drivers identification card is used, the rich vehicle owner may still get discounted prices.  Unless the system links a car with the driver.... The system gets complicated as you can see.  But it will probably also result in black market sales of cheap diesel from one side to the other.

You could probably study what Malaysia has done, and see the effects to see what will be the end result. 

Land use will depend on incentives given by the government to investors.  If for instance, the area has some resource, and the nation wishes to exploit that, it could give tax holidays and other incentives to attract the investor to convert that raw material.

In many cases, fuel cost may not be so significant a deterrent, unless you are in countries like China and India, where inland transportation costs are a significant element in total cost of the product.

When you compete domestically on an industrial basis, whatever the fuel price may be, if all manufacturers enjoy the same fuel price, then there is no deterrent to investment.  Hence land will be used by such industries and preference will be close to urban areas in order to have access to employees and markets for their products.  Generally infrastructure is also better nearer to urban areas.

Main users of fuel will be transportation followed by industries generally. 

Not to my knowledge, and also this is not my area of research, mine is soil parameter.

Thank you!  Guy and Nina, the papers you suggested are very useful.

Yes there is some studies which try to estimate evapotranspiration using precipitation and land cover (not necessarily based on CORINE classification). eg. is the study done by Ward E. Sanford and David L. Selnick

ESTIMATION OF EVAPOTRANSPIRATION ACROSS THE CONTERMINOUS UNITED

STATES USING A REGRESSION WITH CLIMATE AND LAND-COVER DATA

in JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION

where the authors relate precipitation and land cover to estimate evapotranspiration in the following way (see attached file)

Basically, two ways:empirical way and model way.

The first uses empirical relationships between Vegetation Indices and biomass to estimate large area biomass from remote sensing imagery.

The second one uses RS observation as input to drive ecology model to predict biomass.

Try Plugins/Segmentation/Advanced Weka Segmentation or Trainable Weka Segmentation (plugin name was changed recently, so it depends on your Fiji version).

Manual can be found here: http://fiji.sc/Trainable_Weka_Segmentation, but for my images it worked pretty well without changing anything - just train the segmentation by drawing lines on your image and assign them to respective classes (you can have more than two classes). Train the segmentation, save it, and then run it on different images to see if it performs well.

The segmentation output is a grayscale image with shades of gray representing different classes. You can measure the areas by setting threshold (Ctrl+Shift+T) and execute Measurement (Ctrl+M). Make sure that you have Limit to threshold option set to on in Analyze/Set Measurements...

Hope that helps,

Ales

Avit Bhoumick has shown you a easier way but it is academic, useful in completing an academic project. Kasper has shown you a right way if you intend to predict really. I am showing you the way assuming that you are on task of planning and want to reach implementable conclusion.

Firstly future is unpredictable. In history we do not find any example where the prediction was done in land use simply because transitions depend upon numerous variables of development in future. e.g. Govt. policies to people's attitude to technological development. YET PREDICTION IS MANDATORY FOR PLANNING THE FUTURE.

Hence the best way is to try to study THE PRESENT TREND INSTEAD OF TRYING TO SURVEY THE PAST TREND, that does not really help. Primarily because the trends do also change, forward backward, left right. One should take up a huge survey and opinion collection of a huge no. of people from various walks of life. PARTICULARLY NEW GENERATION. Their aspirations, expectations, direction of thoughts, govt. plans, Success and failure of plans, Corporate sector, International scenario, Agriculture, expansion of urban areas, whether the migration may reverse (Strong reversal trends are seen in some regions where rural development in form of rural industries is taking place fast) and what not. IT IS A HUGE TASK. But in my opinion there is no other realist way. To collect information, to find a futuristic trend out of it, predict periods, directions, scale etc. is further difficult and should be done with collective analysis. (Not by a single brain) You will have to live long to see the extent of success of your planning done on the basis of these prediction. EVEN YOU SUCCEED 10% YOU SHOULD PAT YOURSELF.