Integrating GIS and Numeric Weather Prediction Model with Wheat

Simulation Model for Optimal Wheat Production Locations in

Arid Regions

R. Al-Habsi

, Y. A. Al-Mulla

, Y. Charabi

, H. Al-Busaidi

and M. Al-Belushi

Department of Soils, Water and Agricultural Engineering, College of Agricultural and Marine Sciences,

Sultan Qaboos University, P.O.Box 34, Al-Khod 123, Muscat, Sultanate of Oman

Department of Geography, College of Arts and Social Sciences, P.O.Box 42, Al-Khod 123, Sultan Qaboos University,

Muscat, Sultanate of Oman

yalmula@squ.edu.om

Keywords: Numeric Weather Prediction Model, Simulation Model, Wheat, Emergence, Geographic Information

System, Arid Regions.

Abstract: The upgrade rate of self-sufficiency in wheat depends largely on the amount of water and land to achieve

the quantity and proportion of self-sufficiency target. The climatic and soil conditions, however, are

dynamic conditions. Hence, these conditions seriously limit the capability of providing the optimum tempo-

spatial required data to assist in improving the wheat production unless specialized sensors are utilized

along with excessive work. That implies the crucial need of using computer simulation models. The general

objective is of this study was to delineate the best location for wheat production in arid regions such as

Oman through linking Wheat Simulation Model (WSM) with Numeric Weather Prediction Model (NWPM)

in Geographical Information Systems (GIS). The GIS application software used in this study was the ESRI

ArcGIS. Four field trials, over two seasons, have validated positively the linkage of the developed WSM

with GIS. The developed model can be promoted as a tool of improving wheat cultivation through making

the most of available water in wheat production and increasing the growing acreage of wheat in arid regions

like Oman.

1 INTRODUCTION

Food security is an important issue especially after

the recent food crisis that hit many parts of the

world. Food crisis was not limited only to the

substantial rise in the prices of food imports, but

extended to the lack of universal availability, and

then the scarcity and the difficulty of obtaining these

goods. One of the most important crops for food in

the world is wheat, where more than two thirds of

food is provided by cereals (Bushuk and Rasper,

1994) and one third production of cereals is wheat

(Carver, 2009). Knowing when to plant wheat crop

is one of the most important factors for better

emergence timing which leads to better wheat yield

especially in arid regions. The emergence time,

however, is affected by climatic conditions, soil

property and planting depth (Al-Mulla et al., 2014).

Moreover, despite the fact that an arid country like

Oman has no problem with soil temperature for

wheat emergence, Omani wheat growers usually

sow almost double number of seeds into a land in

order to assure they will get higher wheat production

efficiency of that land due to non-emergence of

portion of the sowed seeds. That might due to seeds

viability and to soil properties especially the

optimum soil temperature and soil water potential in

addition to planting depth which all play an

important role in determining the time of emergence

for wheat seedlings. Therefore, in order to increase

the wheat production efficiency, there is a need to

determine the optimum soil temperature, water

potential and planting depth of wheat and so the best

time and percentage of wheat emergence can be

predicated. However, given the above constraining

factors and due to varying values of these factors

from time to time and from place to place,

estimation of emergence time for wheat will be

much easier by using a computer simulation model.

Simulation is a computerized model that

describes the behavior of a complex system based on

141

Al-Habsi R., Al-Mulla Y., Charabi Y., Al-Busaidi H. and Al-Belushi M..

Integrating GIS and Numeric Weather Prediction Model with Wheat Simulation Model for Optimal Wheat Production Locations in Arid Regions.

DOI: 10.5220/0005472501410147

In Proceedings of the 1st International Conference on Geographical Information Systems Theory, Applications and Management (GISTAM-2015), pages

141-147

ISBN: 978-989-758-099-4

 2015 SCITEPRESS (Science and Technology Publications, Lda.)

a set of data and dynamic variables and interactive

components. The Computer simulation models

become popular in many natural systems in physics,

chemistry and biology, human systems in economics

and social science. Biosystems field is considered as

one that requires the use of simulation and computer

modeling including the predicting the time and

percentage of wheat emergence in field applications.

A wheat simulation model (WSM) was

developed by Al-Mulla et al., (2014) to predict the

time and percentage of winter wheat emergence

based on the three above mentioned factors: planting

depth, soil temperature, and soil water potential. The

developed wheat emergence model is based on the

hydrothermal time concept which was proposed and

further developed, but for germination only, by

Gummerson (1986), Bradford (1990 and 1995) and

Cheng and Bradford (1999). The WSM governing

equation has the following form:

)])(()[(

)(

fprobit











(1)

where t(f)E is time from sowing seeds to emergence

(days),  is soil water potential (MPa), and



and

σψb are mean and standard deviation of the base

water potential respectively (MPa), T is soil

temperature (°C), Tb is the base soil temperature

(°C), and probit(f) indicates the number of standard

deviations away from the mean that any fraction of

the seed population lies. It linearizes a cumulative

normal distribution which facilitates modeling

efforts (Bauer et al., 1998). The parameter HT is

the hydrothermal time to emergence (MPa degree-

days). It can be determined using the following

formula:











(2)

where all parameters explained above.

The parameter probit(f) can be determined as

follows:



Eprobit











)50(

)(





(3)

where

)50(



is base water potential for 50%

population of planted seeds

The developed WSM also includes the

calculation of the maximum percent of emergence

(Emax) by using the following equation which

works as a threshold for the developed model:

max

cDcDcE 

(4)

where D is sowing depth (cm), and c1, c2, and c3

are constants related to ψ and T.

The soil temperature and water potential

information required for the emergence model are

not easy to find for any lands. That’s due to the need

of using special sensors to measure these factors at

time of emergence of the wheat crop. In this study a

novel approach is explored in order to utilize the

WSM model without the need of using any sensors

by finding a better way for extracting the needed

model’s input data from Numerical Weather

Prediction Model (NWPM) which is based on

numerical models that deal with set of motion

equations. These equations govern the fluid flow

partial differential equations (Stensrud, 2007).

Hence, by linking WSM with NWPM in Geographic

Information System (GIS), the best location of

wheat production in arid regions as the Sultanate of

Oman can be delineated and that was the main aim

of this study. The specific objectives of the projects

were: 1) to validate the developed simulation model

for predicting the time and percentage of emergence

using local field data, 2) to examine the performance

of the model using different wheat cultivars, 3) to

extract needed data from Numerical Weather

Prediction Model (NWPM), 4) to re-design the

simulation model by linking it with Geographic

Information system (GIS) and NWPM, and 5) to

create maps delineating the best location of wheat

production in the Sultanate of Oman.

2 MATERIALS AND METHODS

2.1 Field Experiment

Field experiment was conducted at Agriculture

Extension Station in Sultan Qaboos University in the

Sultanate of Oman (Figure 1) between first of

December 2013 until End of May 2014. The field

work was divided into two times. In time one,

planting started in 1st December 2013 until the end

of harvesting stage March 2014. Planting in second

time started on 22 January 2014 until 21 May 2014.

The field was divided to four lines, the first two lines

for time one and the last two lines for time two.

Each line was divided to seven 2m ˟ 2m plots. Each

plot was divided to ten cultivated lines irrigated by

drip irrigation system. The space between one line to

another was 20 cm and the wheat seeds were sowed

adjacent to the lines with 5cm spacing. Two wheat

varieties were used. One of them is local variety

GISTAM2015-1stInternationalConferenceonGeographicalInformationSystemsTheory,ApplicationsandManagement

142

which is Coli and the second one was imported from

Kuwait which is KW1. The seeds were sowed at two

different depths: 2.5cm and 5cm. Hence, the

experiment treatment factors are: 2 times ˟ 2

varieties ˟ 2 planting depths. Three replicates for

each treatments planting with one control plot of

each varieties were conducted. The planting method

in the control plots was by broadcasting.

Fertilization was applied as recommend by AES

staff experience. By which fertilizer was applied two

times a week through fertigation system where

combination of urea, potassium nitrate and

phosphoric acid was applied trough the irrigation

system.

2.2 Germination Test

Ten petri dishes were prepared for each variety and

in each petri dish a filter paper and ten seeds from

each variety were putted and irrigated with water.

So, 100 seeds in total for each variety were

examined. Germination of seeds was counted and

recorded on daily basis.

2.3 Physical and Chemical Soil

Analysis

2.3.1 Soil Sampling

One composite soil sample was taken by collecting

different soil samples from different plots and mixed

together. Then soil was air dried and sieved through

2 mm.

Figure 1: Sultanate of Oman (Source: geoatlas.com).

2.3.2 Moisture Content and Oven Dry

Weight

After the soil sample was air _ dried and sieved by

2mm sieve, unspecified amount of soil was taken

and its weight was determined before and after

drying in oven at 104 °C for 24 hours. Then

percentage of moisture content and Oven dry weight

were calculated and result used in soil texture

calculation.

2.3.3 Soil pH, EC and SAR Measurements

Paste saturation sample was prepared and soil

solution was extracted by air vacuum for Electrical

Conductivity (EC) and pH measurements using

electrical electrode. For Sodium Absorption Ratio

(SAR), concentration of Sodium (Na), Calcium (Ca),

and Magnesium (Mg) were measured using ICP. To

find SAR value, the following equation was used:

 





  

(5)

2.3.4 Soil Texture

Soil sample was air-dried and sieved by 2mm sieve.

From sieved soil, 50g of soil was assigned into a

baffled stirring cup with 10ml of 0.5N sodium

hexametaphosphate and distilled water was added

until half fill of the cup. The mixture was stirred for

five minutes and transferred to 1000ml graduated

cylinder which was filled with distilled water until

1000ml mark. The suspension was mixed and at the

end of 20 second from mixing, the hydrometer was

inserted. The first reading of hydrometer was

recorded after 40 seconds and also temperature of

suspension was recorded by the thermometer. Then,

the hydrometer was removed and re-shacked again.

After two hours, the reading of hydrometer and

temperature were recorded. The percentage of each

particle size was calculated according to these

equations:

Corrected Hydrometer Reading:

HRc = HR + (0.2 for every 1°C above 20°C)

(6)

% 

  40 sec  2



∗100

(7)

% 

  2



∗ 100

(8)

%  100 – %  %

(9)

IntegratingGISandNumericWeatherPredictionModelwithWheatSimulationModelforOptimalWheatProduction

LocationsinAridRegions

143

2.4 Irrigation

Drip irrigation was used in the study. Ten drip lines

were crossing the plots with 20cm spacing between

the emitters. Water required was calculated based on

crop evapotranspiration for each stage of wheat

development. The historical data like daily

maximum and minimum soil temperature and

humidity, wind speed and the meteorological data

for Seeb weather station were used. Visual basic in

Excel sheet was used to create sheet for reference

water requirement (ETo). The sheet is based on

Penman – Monteith method. To determine the time

of irrigation, the discharge of water from meter was

recorded. Hence, the time of irrigation was

calculated by dividing volume of water required by

the discharge.

2.5 Sensors

The inputs parameters including water potential and

soil temperature, for wheat simulation model can be

obtained from the sensors installed in the field. A

229 heat dissipation sensor from Campbell Scientific

Company, USA, is used to measure water potential

indirectly using principle of heat dissipation. The

principle of heat dissipation is whenever there is

water potential gradient between sensors and the

surrounding soil, water movement between sensor

and soil take certain time to reach equilibrium.

Hydraulic Equilibration time depend on magnitude

of water potential gradient and hydraulic

conductivity. The changes in water content of sensor

ceramic matric lead to change in the thermal

conductivity of sensor/ soil complex. There is

exponential relationship between water content and

thermal conductivity. As water content in the

ceramic sensor increase, the thermal conductivity

increase. A 229 heat dissipation sensor is porous

ceramic cylindrical shaped with thermocouple and

heating element at the middle of the cylinder. It has

the ability to measure a wide range of matric

potential from -10 to -2500KPa and it is compatible

with most Campbell Scientific data logger and

multiplexer. Also, it is known by long lasting

without need for maintenance. The 229 should be

installed horizontally at the desired depth and good

contact between the ceramic cylinder and soil must

be exist (Instruction manual of model 229, Campbell

Scientific Inc.). Soil water potential can also be

measured by using MPS2 from Decagon Devices

Inc., USA. MPS2 is ideal sensor for a range of water

potential measurement between -0.01 and -0.5 MPa

and soil temperature between -40 and 60C (Decagon

Devices Inc.). A 5TE sensor from Decagon devices,

USA, is used to measure volumetric water content,

soil temperature and electrical conductivity. It

measures the three parameters independently. The

volumetric water content is obtained by measuring

the dielectric constant of the media using

electromagnetic field supplied from the sensor with

70 MHz while the soil temperature is obtained from

thermistor which is installed at surface of the sensor

with a range of readings of -40 to 50°C. Electrical

conductivity is obtained by using stainless steel

electrode array and the reading is taking within the

range of 0 to 23 ds/m. The sensor is easily installed

in the field through pushing it directly to

undisturbed. EC is measured by applying alternating

current to two electrodes and measuring the

resistance between them soil (Decagon Devices

Inc.).

2.6 Linking WSM to GIS

To link the wheat simulation model in GIS, soil

temperature and volumetric water content raster

layers were required as input for the model and they

were extracted from the European Centre for

Medium-Range Weather Forecasts (ECMWF)

website which is an independent intergovernmental

organization established in 1975. The Centre

provides a catalogue of forecast data worldwide that

can be purchased by businesses and other

commercial customers for national community.

Different spatial analysis tools in GIS were used to

create the final emergence map mainly rater

calculator and reclassifying tools (Figure 2).

2.7 Statistical Analysis of Physical

Characteristics of Wheat

Physical characteristics of wheat like length of plant,

number of spikes per plant, number of seeds per

spike and weight of 100 seeds of each variety were

examined at the late season of each planting time.

Plant height was measured by taking average length

of random selection of plants using meter tab from

the soil surface to the begging of the spike. The

average plant length in each plot was recorded.

Random plants were also selected to find out

number of spikes in it and average numbers of seed

in each spike were counted. After the seeds removed

from its spike, 100 seeds from each plot was

weighed using digital scale. For statistical analysis,

ANOVA for single factor was used to examine the

differences among each factor.

GISTAM2015-1stInternationalConferenceonGeographicalInformationSystemsTheory,ApplicationsandManagement

144

Figure 2: Flowchart of Creating Wheat Prediction Model

in GIS.

3 RESULTS AND DISCUSSION

Germination was started after second day of

planting. For local variety (Coli) germination was

99% in second day and 100% in third day. For

KW1, germination was 46% in second day and after

second day till day six of planting, germination was

stopped at 60%.

Soil Texture Class for planting field is silt loam

which is medium textured soil. For wheat

production, the best soil type is well drained fertile

loamy soil to sandy loam soil (DAFF, 2010).

Electrical conductivity for paste soil extract is 478

µS\cm, Sodium absorption ratio (SAR) is 2.35 and

pH is 8. The best range of soil pH for wheat

production is between 6 and 7.5 (DAFF, 2010). The

pH value in this study exceeded the optimum range a

little bit. Alkalinity affects wheat production

negatively in two ways by decreasing water

infiltration and decreasing availability of

micronutrients. Since that the solubility of Iron (Fe),

manganese (Mn), copper (Cu) and zinc (Zn)

decrease with increasing soil pH. However,

deficiency of the last three ions due to alkalinity has

not severe effect in wheat production area (Heyne,

1987). Based on the EC and SAR values, soil was

classified as non-saline and non sodic soil. It was

normal soil and suitable for wheat production.

Crop evapotranspiration and Irrigation

Scheduling for each plant stage are presented in

Table 1 for planting time one and two. As expected,

for first time of planting, crop coefficient (Kc)

started with small value but it increased as plant

grown and developed and then decreased at late

season as plant senesce. The crop water requirement

(ETc) followed the same pattern of the Kc. Because

the plant need more water at growth and developing

stages, the duration of irrigation increased until the

late season where it decreased and stopped for one

week before harvesting. The ET₀ and ETc values

increased during time two of planting than time one

due to the increase of air temperature. That resulted

in increasing the duration of irrigation supply in time

tow more than in time one.

Since WSM requires soil water potential as one

of the input data and since both sensors the 5TE and

MPS-2 measure volumetric water content for, there

was a need to convert these sensors data into water

potential. A relationship between soil’s volumetric

water content and water potential were obtained. By

which a regression equation of calculating water

potential from volumetric water content was found

as follows:

  20.208 



 11.747  1.7176

(10)

Where Ψ is soil water potential (MPa) and wc is soil

water content (%).

The time of emergence equation of the WSM has

a probit parameter. It is a value that is used to

indicate the number of standard deviation away from

the mean that any fraction of seed population lies.

Also, it is used to linearize a cumulative normal

distribution which makes the model easy to work

(Al Mulla et al., 2014; Bauer et al., 1998). Since it is

not possible to calculate the probit value in GIS

directly, relationship was found between emergence

percentage and the probit value which can be

expressed in the following equation:

  0.3383 ln







 1.996

(11)

where Probit is probit value and E is emergence

percentage (%).

Figure 3 shows emergence results from time one.

Emergence started after day 4 from planting for all

treatments and reached maximum emergence at day

12 after planting. The maximum percentage was

achieved for Coli variety at 2.5 cm planting depth

which was 75 % at day 12 after planting while at

planting 5 cm depth, the maximum emergence was

62 %. For KW variety, at planting depth 2.5 cm, the

maximum emergence was achieved after day 12

with 53 % and at planting depth of 5 cm, the

maximum emergence was 36 %. Emergence results

of time 2 are presented in figure 4. The emergence

of both varieties started after day 5 from planting

and reached the maximum at day 12 from planting.

Coli variety had higher percentage of emergence

than KW variety where at 2.5cm planting depth it

reached the highest percentage among other

treatments with 77%. The KW variety reached its

higher emergence percentage at shallow planting

depth (50%) in comparison to deep planting depth

(44%).

The WSM predictions for the wheat to start

emerging were 5 days after planting (DAP) in time

one and 5 DAP in time two. While in the field it

IntegratingGISandNumericWeatherPredictionModelwithWheatSimulationModelforOptimalWheatProduction

LocationsinAridRegions

145

Table 1: Evapotranspiration and Irrigation Scheduling for Times one and two.

took 4 DAP in time one and 5 DAP in time two. The

DAP to reach 50% emergence was predicted by

WSM as 6 DAP and 6 DAP in both times one and

two. In the field, the wheat reached 50% emergence

after 7 DAP and 6 DAP in time one and time two.

The maximum percentage achieved in the field, for

time 1, was 74% whereas WS

prediction was 67%

while for time 2, the maximum percentage was

achieved in the field was 77% whereas WSM

prediction was 80%. The DAPs to reach maximum

emergence was predicted by the WSM as 7 DAP

while in the field it was achieved after 11 DAP for

time 1. For time 2, the SWM prediction was 6 DAP

in the field it was achieved after two more days

which was 8 DAP.

Figure 3: Wheat seed Emergence for Time one.

Figure 4: Wheat seed Emergence for Time two M.

The soil temperature and volumetric water content

data at level 1were extracted from ECMWF by

selecting a month long (October 2013) of the

required data with time steps of 24 hours. Equation

12 was used in the raster calculator to find soil water

potential from the volumetric water content layer

extracted from ECMWF. The data format was in

Network Common Data Form (NetCDF), so they

needed to be converted to raster data format for GIS.

That was done by selecting “Make NetCDF Raster

Layer” which is a sub menu under Multidimintion

menu of ArcToolbox.

After raster layers of soil temperature and water

potential were created from NetCDF file, coding the

equations of hydrothermal time and maximum

emergence was done using raster calculator.

Equations of maximum emergences are based on

certain range of temperature and water potential with

a total of 23 conditions. Hence, 23 raster layers of

pixels that match with emergence conditions have

certain values of maximum emergences and

therefore they made to appear in colored pixel.

However, the other pixels that do not match the

emergence conditions of emergence have no data.

To merge all emergence raster layers in one layer to

be used in the final equation of time of emergence,

the no data pixels were converted to zero value using

Reclassify tools from spatial analysis. All

reclassified emergence raster layers were combined

in one raster layer using addition tool in raster

calculator producing a single layer map. Figure 5

shows the outcome of the GIS-Linked WSM

modified model for the Arabian Peninsula region

after extracting the NWMP data from the ECMWF.

4 CONCLUSIONS

Wheat is one of the most important crops for food in

the world including arid regions as the Sultanate of

Oman. For arid regions, It would be especially

useful to develop a method that can increase wheat

production location but with minimum possible

water consumption. By integrating GIS technology,

a Wheat Simulation Model (WSM) was further

developed for optimal wheat production location in

arid regions. WSM was tested with field data and the

effectiveness of the model was proven. The study

Satge Duration

Average ET

₀

Average Etc Volume Irrigation Time

Stage Name Days mm/day mm/day

min

Time 1 Time 2 Time 1 Time 2 Time 1 Time 2 Time 1 Time 2 Time 1 Time 2 Time 1 Time 2

Initial Stage 20 20 0.4 0.4 3.3 3.6 1.32 1.4 0.07 0.08 1.3 1.4

Growth and Branching 30 30 1 1 3.1 4.5 3.1 4.5 0.17 0.25 3.1 4.6

Completion of growth and Flowering 30 30 1.2 1.2 3.5 5.4 4.2 6.5 0.24 0.36 4.3 6.7

Compos itio n and Grain Filling 30 30 1 1 4.7 6.7 4.7 6.8 0.26 0.38 4.8 7

Late Season 15 15 0.4 0.4 5.3 8.8 2.12 3.5 0.12 0.20 2.1 3.6

GISTAM2015-1stInternationalConferenceonGeographicalInformationSystemsTheory,ApplicationsandManagement

146

illustrates a good example of GIS application for

solving spatial and temporal crop production

optimization problem. By which, it becomes

possible to delineate (map) the WSM outcome that

can cover any part of world including arid regions

by eliminating the need of using any sensors to run

the simulation.

(a) (b)

Figure 5: GIS linked WSM outcome: (a) soil Water

Potential, (b) Soil temperature, (c) Hydrothermal time

factor and (d) Time to emergence.

ACKNOWLEDGEMENTS

The corresponding author, as a PI, appreciates the

financial support for this research provided by

Sultan Qaboos University (IG/AGR/SWAE/12/02).

REFERENCES

Al-Mulla, Y. A., Huggins, D. R., Stöckle, C. O., 2014.

Modeling the Emergence of Winter Wheat in

Response to Soil Temperature, Water Potential and

Planting Depth. Transactions of the ASABE, 57(3):

761-775.

Bauer, M. C., Meyer, S. E., Allen, P. S., 1998. A

simulation model to predict seed dormancy loss in the

field for Bromus tectorum L. J. Exp. Bot. 49:1235-

1244.

Bradford, K. J., 1990. A water relations analysis of seed

germination rates. Plant phys. 94:840-849.

Bradford, K. J., 1995. Water relations in seed germination.

In: Seed Development and Germination. Marcel

Dekker Inc. New York: 351-396.

Bushuk, W., Rasper, V. F., 1994. Wheat production,

properties and quality. Blakie Academic and

professional an imprint Chapman and Hall 273-310.

Carver, B. F., 2009. Wheat Science and Trade. World

Agriculture Series. Wiley-Blackwell. Ames, Iowa, 1

edition.

Cheng, Z., Bradford, K. J., 1999. Hydrothermal time

analysis of tomato seed germination responses to

priming treatments. J. Exp. Bot. 50:89–99.

DAFF (Department of Agriculture, Forestry and

Fisheries). 2010. Wheat Production Guideline, Public

of South Africa.

Gummerson, R. J., 1986. The effect of constant

temperatures and osmotic potentials on the

germination of sugar beet. J. Exp. Bot. 37, 729-741.

Heyne, E. G., 1987. Wheat and Wheat Improvement,

Second Edition, Publisher: Madison, Wisconsin, USA.

MAF (Ministry of Agriculture and Fisheries)., 2009.

Wheat Production in the Sultanate of Oman: Reality,

Challenges, and Future Vision. Wheat Production

Project National Committee Report. SQU. Oman.

Stensrud, D., 2007. Parameterization Schemes Keys to

Understand Numerical Weather Prediction Model,

Cambridge.

IntegratingGISandNumericWeatherPredictionModelwithWheatSimulationModelforOptimalWheatProduction

LocationsinAridRegions

147