INTRODUCTION
The magnitude of human influence on natural ecosystems usually ends with
widespread environmental problems such as soil erosion, floods and droughts threatening
human being. Rangeland deterioration (and erosion) has accelerated in current
decades, primarily due to a doubling or tripling of livestock numbers, extensive
plowing of rangelands, firewood cutting, expansion of well drilling into formerly
inaccessible areas, and better transportation facilities. The area of abandoned
arable land in Iran has doubled in recent years and the number of livestock on
grazing lands is estimated to be two to three times the carrying capacity. The
prevention of soil erosion, which means reducing the rate of soil erosion to approximately
that which would occur under natural conditions, relies on selecting appropriate
strategies for soil conservation (
Morgan, 1979).
Although it is impossible to stop soil erosion completely under natural
conditions, there is a great need to control erosion for proper land and
water use planning. This requires awareness of soil erosion and foreseeing
changes such as in land use.
On a global scale the annual loss of 75 billion tons of soil costs the world
about US$400 billion per year, or approximately US$70 per person per year (Eswaran
et al., 2001). There is no official document on the amount of soil
erosion in Iran. Based on a research on the suspended sediment data of more
than 200 sampling stations around the country, average annual suspended sediment
yield is reported to be 2 t ha-1 or 350 million tons (Arabkhedri,
2003). Assuming Sediment Delivery Ratio (SDR) of 17.1 to 21.6 (Ouyang
and Bartholic, 1997) and the amount of bed load to be 20% of the amount
of suspended load, the amount of soil erosion in Iran is some 2 billion tons
(2.7% of the world’s soil loss). Perhaps one of the most costly results of soil
erosion is related to damage done by the soil particles that are dislodged and
moved downwind or downstream. Sedimentation raises streambeds, reducing the
depth and capacity of the channels. Sedimentation of lakes and reservoirs reduces
their capacity, value and life expectancy (Frederick et
al., 2003). Each year, 550 Mm3 of new dam reservoirs are
built in Iran from which more than 200 Mm3 is filled by sedimentation
(Samadi Broujeni and Shamsaei, 2007). Erosion has become
an environmental problem (Ananda and Herath, 2003) as well that must be remedied
for the sake of clean air and water. Soil particles adsorb pollutants such as
pesticides, fertilizers and different industrial and municipal chemicals that
are best kept out of water by keeping the soil on the land (Foster,
1988; Wanielista and Yousef, 1993). It has therefore
economic, political, social and environmental consequences due to both on-site
and off-site damages caused by soil erosion.
The conflict between environmental protection and the economic issues are challenges
facing land use planners and decision makers in many developing countries (Chang
et al., 1995; Gezelius and Refsgaard, 2007).
Proper environmental planning needs especial consideration on land use scenarios
and optimization. The watershed optimization for each land use, especially agriculture
as one of the significant contributors to the environmental degradation, is
therefore necessary to achieve sustainable development (Seppelt
and Voinov, 2002; Heilman et al., 2003; Wang
et al., 2004).
Nikkami et al. (2002) indicated that land use
optimization is one of the appropriate strategies for soil conservation. They
used land use optimization for minimizing soil erosion and maximizing farm production
of each land use in Damavand watershed, Iran. The expected annual soil erosion
from the entire watershed was reduced by 5% and the annual net farm income was
increased by 134%. Mohseni Saravi et al. (2003)
used goal programming in Garmabdasht watershed in Golestan province, Iran, to
determine the optimal solution for different activities in the watershed. Industrial
forest, pasture, park and protected areas were optimized with the goals of maximization
of benefit, production, employment opportunities, and minimization of total
investment and sediment yield. Kralisch et al. (2003)
and Riedel (2003) combined artificial neural network
and GIS with LP to maximize benefits gained from land utilization in a watershed
in Germany and mountainous area of North Thiland, respectively.
Benli and Kodal (2003) developed linear and non-linear
programming models in South-east Anatolian watershed in Turkey for determination
of optimum cropping pattern, water amount and farm income under two scenarios
of adequate and limited water supply conditions. Wang et
al. (2004) used LP and GIS for land use optimization based on existing
land use, slope, distance to surface water and conversion preferences in Lake
Erhai basin, China. Luo and You (2007) presented a modeling
approach to investigate water quality trading in soil erosion control, based
on watershed simulation and optimization models in which various uncertainties
were reflected within the Swift Current Creek watershed, Canada. Sadeghi
et al. (2008) used a multi-objectives linear optimization problem
for minimizing soil erosion and maximizing farm production of each land use
in Brimvand watershed, Iran. The results of the study revealed that the amount
of soil erosion and benefit could respectively, reduce and increase to the tune
of 7.9 and 18.6%, in case of implementing optimal allocation of the study land
uses.
Considering scarcely documented researches in land use management and
protecting watershed resources applying optimization approaches, the present
study has been conducted to optimize land resources allocation to orchard,
rangeland, irrigated cropland and dry farming within the Kharestan watershed
in the northwest of Eghlid city, Fars province, Iran, using a multi-objective
linear programming approach.
MATERIALS AND METHODS
This study was conducted in Kharestan watershed located in upstream of
Doroodzan Dam in the north west of Eghlid city in Fars province, Iran,
during years 2006-2007. It extends between 30°E35’ to 30°E47’
N latitude and 51°E47’ to 52°E00’ E longitude and covers an area
of 14685 ha (Fig. 1).
The average yearly precipitation is 580 mm in a Mediterranean and semi-wet
climate condition. Maximum, minimum and average elevations are 3040, 1900 and
2337 m above sea level and average land slope is 25.67%. The information and
data required for computation of soil erosion and net income in each land use,
land and water availability, soil characteristics, land slope and socio-economical
conditions were extracted from the available studies of Fars province Watershed
Management Office in addition to some other field studies and land surveysfor
further details and information. The slope, land component, land use and erosion
maps as a part of necessary maps in the study area have been shown in Fig.
2.
 |
| Fig. 1: |
Location of the study area on the Iran map |
Kharestan contains a diversity of land uses and slope classes. The major
land uses are rangeland, dry farming, irrigated farming and orchard with
the areas of 10550, 1050, 871 and 561 ha, respectively. Based on linearity
of objective functions, multi-objective linear programming model was applied
for three different scenarios of land use combination and land management.
Scenario 1: Existing land uses plus land management, to show the
effect of land use optimization with no any change on the land management
practices
Scenario 2: Existing land uses with some degree of land management, to show
the effect of very simple land management activities.
Scenario 3: Proper land uses plus land management, to show the effect of both
land use optimization and land management on minimizing soil erosion
and maximizing net income.
In the first scenario, existing land uses were mapped using 2002 Landsat imagery
and checked by field work. With no change in the area of existing land use,
land management practices were applied on these areas within the second scenario.
In the last scenario, all scientific aspects of land suitability (Mahler,
1979; Brengle, 1982) and land use management were considered.
The amount of soil loss in each land use is estimated from the application
of modified Pacific South-west Inter-Agency Committee model (Johnson
and Gebhardt, 1982) in ILWIS-GIS and applying the concept of sediment delivery
ratio. For proper management of agricultural lands, it is not wise to have dry
farming on slopes greater than 12% and irrigated farms on slopes greater than
5%. Recommended slopes are milder than these slopes in order to avoid soil erosion
and reduction of crop yield.
All benefit/cost data of the crops were collected through field studies.
Major orchard crops that are included in this model were apple, walnut,
egg-plum, peach and almond. Irrigated farms were planted with wheat, barley
alfalfa and cucurbit. Dry farming crops were wheat, barley, and lentil.
The weighted average dry-forage production, their Total Digestible Nutrients
(TDN) and total animal units per hectare were also determined.
There is no research on the evaluation of economic losses due to soil
erosion in the study area. Therefore, it is difficult to evaluate it directly.
However, these losses can be estimated indirectly by the evaluation of
fertile soil loss. For example, based on data relating topsoil loss to
yield reduction, just 2.5 cm of topsoil loss is sufficient to reduce U.S.
wheat yields by an average of 60 million bushels (bushel = 35.21L) per
year. Another way to estimate economical losses due to soil erosion is
to apply lost soil to the eroded area based on the depth of root zone
in each land use.
 |
| Fig. 2: |
Slope (top left), land components (top right), land
use (bottom left) and erosion (bottom right) maps for determination
of allocable land resources within Kharestan watershed, Iran |
The depth of the lost soil in each land use is calculated by considering the
amount of soil erosion in that land use, the appropriate rooting depth of vegetation
(root zone) and soil bulk density. The general form of a multi-objective optimization
problem in the Kharestan watershed with n decision variables, m constraints
and p objectives to minimize soil erosion and maximize net income is as given
in Eq. 1-4 (Nikkami et al., 2002).
Subject to:
where, Z1 and Z2 are the annual net farm income
in million Iranian Rails (mIR) and the total annual soil loss (t), respectively.
In each land use Xi, Ci, Ai1,Ai2
and Ai3 are surface area (ha), annual soil loss
per unit area (t ha-1), amount of net farm income (mIR ha-1),
production cost (mIR ha-1) and cost due to soil loss (mIR ha-1),
respectively. B is the total land area (ha-1). The problem
can be written in detail in the following form:
Subject to:
where, X1 through X4 are areas allocated to orchard,
rangeland, irrigated farming and dry farming (ha), respectively. A11
through A41 are amounts of net farm income per unit
area of orchard, rangeland, irrigated farming and dry farming (mIR ha-1).
A12 through A42 are production
costs per unit area of orchard, rangeland, irrigated farming and dry farming
(mIR ha-1). A13 through A43
are erosion costs per unit area of orchard, irrigated farming and
dry farming (mIR ha-1). C1 through C4
are annual soil loss per unit area of orchard, rangeland, irrigated farminghj
and dry farming (t ha-1). B1 through B7
are maximum limits of orchard surface area, surface area of irrigated
farming, surface area of dry farming, surface area of orchard plus irrigated
farming, total area, lower limit of orchard surface area, and surface
area of rangeland (ha), respectively. There are 25 springs with discharges
from 1 to 30 L sec-1 and two rivers of Kharestan and Tizab
located in the west and east of the watershed, respectively with annual
discharge of 8.53 Mm3 that is sufficient for agricultural development.
Therefore, no constraint was defined for water availability.
Due to not being able to make any changes in the use of urban lands,
these areas were excluded from land use optimization.
RESULTS AND DISCUSSION
The solution of optimization procedure for each scenario is given below.
First scenario: The general form of the optimization problem is
written as follow.
By simplifying the first objective function and changing the minimization
to maximization form in the second objective, these equations change to
the following simpler forms.
Second scenario:
By simplifying the first objective function, and changing the minimization
to maximization form in the second objective, these equations change to
the following simpler forms.
Third scenario:
By simplifying the first objective function, and changing the minimization
to maximization form in the second objective, these equations change to
the following simpler forms.
There are eight constraints of the land use optimization model. The constraints
and their justifications are discussed below.
Constraint 1: X1 ≤ 2115
The first constraint indicates that the present area under orchard, which
is 561 ha, could be increased up to 2115 ha. The reason for this constraint
is that the areas of irrigated farms with slope classes of more than 5%
are not suitable for irrigating cropland. These lands could be changed
to other land uses especially orchards, by terracing, if necessary, and
planting permanent vegetation.
Constraint 2: X3 ≤ 237
The second constraint is that irrigated farms, currently 871 ha in area,
after subtracting high slope classes as described in constraint 1, could
not be more than 237 ha.
Constraint 3:X4 ≤ 207
Slopes more than 12% are not suitable for dry farming. The third constraint
indicates that the area under dry farming, which is 1050 ha, after subtracting
high slope classes, could not be more than 207 ha. Other reasons for this
constraint are as follows:
| • |
The government owns the rangelands and people cannot
make any changes. |
| • |
Due to lack of sufficient rainfall in the area, dry farming is not
suitable for most areas in this watershed. |
| • |
People seldom use supporting practice systems in dry farming lands,
which cause large amounts of soil erosion in this form of land use. |
Constraint 4: X1+X3 ≤ 2352
Assuming no limitation on irrigation water, the fourth constraint implies
that the area under orchard and irrigated croplands could not be more
than 2352 ha based on existing slope and soil depth.
Constraint 5: X1+X+X3+X4 =13032
The fifth constraint is simple and it is the area limitation of the Kharestan
watershed after subtracting the urban lands. The sum of the areas under
the four land uses should be equal to 13032 ha of the available lands.
Constraint 6: X1 ≥ 561
As explained in Constraint 1, the sixth constraint forbids reduction
of the present area under orchards.
Constraint 7: X2 ≥ 10550
The seventh constraint indicates that the area under rangeland should
be at least 10550 ha. The reason for this constraint is that the government
owns the rangelands and people cannot change their form of land use (Iran
Forest and Rangeland Nationalization, Act of 56).
Many rangelands have been illegally converted to improper dry farming
lands, which could be changed back to rangelands.
Constraint 8: X1, X2, X3, X4
≥ 0
The last constraint is the non-negative variable declaration. Table
1 shows the area, average annual soil loss and average annual net
income for each land use. Simplified objective functions and the constraints
discussed above for three scenarios i.e, (scenario 1) existing land uses
plus land management, (scenario 2) existing land uses with some degree
of land management and (scenario 3) proper land uses plus land management
are entered in Table 2 to 4, respectively
as revised multi-objective linear simplex tableaus. The computer program
LINGO is used to solve the problems.
After taking allocated areas into account, average annual soil loss and
net income for each scenario is indicated in Table 5 to
7. The results showed that in the optimized condition,
while rangelands experience no change, the area of orchards should be
increased from 561 to 2115 ha (377%), irrigated farms should be reduced
from 871 to 237 ha (73%) and dry farming lands should be decreased from
1050 to 129 ha (88%). In the first scenario the annual soil loss would
have decreased by 4288 t (3.7%) and the annual net income increased by
26,540 mIR (163%). In the second scenario the annual soil loss would have
decreased by 51320 t (37%) and the annual net income increased by 37,360
mIR (206%). In the last scenario the annual soil loss would have decreased
by 71853 t (53%) and the annual net income increased by 37,780 mIR (208%).
| Table 1: |
Area, soil loss, and net income in each land use of Kharestan watershed |
 |
| Table 2: |
Linear multi-objective simplex table of Kharestan watershed in
scenario 1 |
 |
| aColumns 2 through 5 present decision variables,
which in rows 2 and 3 have currency and soil loss units, respectively.
Numbers 1 and 0 in the remaining rows show the presence or absence
of the decision variables in constraints, respectively, bRows
2 and 3 of column 6 indicate the maximization or minimization form
of the objective functions while remaining rows indicate the equality
or inequality form of the constraints, cThe last column
gives the Right Hand Side (RHS) value of each constraint, which represents
land availability in hectares |
| Table 3: |
Linear multi-objective simplex table of Kharestan watershed in scenario
2 |
 |
| aColumns 2 through 5 present decision variables,
which in rows 2 and 3 have currency and soil loss units, respectively.
Numbers 1 and 0 in the remaining rows show the presence or absence
of the decision variables in constraints, respectively, bRows
2 and 3 of column 6 indicate the maximization or minimization form
of the objective functions while remaining rows indicate the equality
or inequality form of the constraints, cThe last column
gives the Right Hand Side (RHS) value of each constraint, which represents
land availability in hectares |
| Table 4: |
Linear multi-objective simplex table of Kharestan watershed in scenario
3 |
 |
| a Columns 2 through 5 present decision variables,
which in rows 2 and 3 have currency and soil loss units, respectively.
Numbers 1 and 0 in the remaining rows show the presence or absence
of the decision variables in constraints, respectively, b
Rows 2 and 3 of column 6 indicate the maximization or minimization
form of the objective functions while remaining rows indicate the
equality or inequality form of the constraints, c The last
column gives the Right Hand Side (RHS) value of each constraint, which
represents land availability in hectares |
| Table 5: |
Land use optimization output of Kharestan watershed in scenario
1 |
 |
| Table 6: |
Land use optimization output of Kharestan watershed in scenario
2 |
 |
| Table 7: |
Land use optimization output of Kharestan watershed in scenario
3 |
 |
Sensitivity analysis of the resource Bi on soil loss and net
income in scenarios 1 to 3 is presented by Fig. 3. In this
figure B1 through B7 are maximum limits of orchard surface area, surface
area of irrigated farming, dry farming, orchard plus irrigated farming,
total area, lower limit of orchard surface area and surface area of rangeland
(ha), respectively. Sensitivity analysis often begins with the investigation
of the effect of changes in the Bi, the amount of resource
i being made available for the activities under consideration. The reason
is that there is generally more flexibility in setting and adjusting these
values than there is for the other parameters of the model.
 |
| Fig. 3: |
Sensitivity analysis: the effect of resources Bi on soil loss (right)
and net income (left) in scenarios 1 to 3 |
The economic interpretation of the dual variables as shadow prices is
extremely useful for deciding which changes should be considered. The
shadow price (yi*) for resource i measures
the marginal value of this resource, that is, the rate at which Z
could be increased by slightly increasing the amount of this resource
being made available. In particular, if yi* >
0, then the optimal solution changes if Bi is changed, so Bi
is a sensitive parameter. Then the investigation continued on Ai1 and
Ci parameters. It was found that , which refers to the restriction of
area under orchard was the most sensitive parameter.
The results approved the applicability of multi-objective optimization
model in solving problems with different objectives which sometimes conflicting
each other. It can also be concluded that the multi-objective linear programming
can be used to tractably search for optimum land use scenarios with respect
to different governing constraints existing within a watershed. The results
also showed the successful linkage between economic aspects and environmental
outcomes at a watershed scale as emphasized by others.
ACKNOWLEDGMENTS
The authors profoundly are grateful to the Soil Conservation and Watershed
Management Research Institute and the Forests, Rangelands and Watershed
Management Organization, both in Tehran, Iran for supplying valuable information
and assistance.
Subscribe Today
