| GISdevelopment.net ---> Application ---> Natural Resource Management |
Infiltration Spatial Variability Map in the Namrood catchment(Tehran, Iran) Base on TM Satellite Image(B3,B4,1998)
M.H. Roohian1 , Prof. Dr. A. M.J.Meijerink2 ,Dr.B.Saghafian3 ,P.sajedian4
1 Scientifitic board member of ANRRC of Booshehr province, Iran,
E mail :roohian@yahoo.com
2 Retired head of water resources department of ITC,Enschede,the Netherlands,
E mail:meijerink@itc.nl
3 Scientifitic board member of SCWMRI of Iran,
E mail : saghafian@scwmri
4 The expert of environmental department Fars province office, Iran,
E mail :sajedian@yahoo.com
ABSTRACT
An infiltration experiment was practiced on the collected field data, during autumn 1999, of a semi arid area, Tehran, Iran to indicate level of site parameters associated with measured infiltration rate and remotely sensed data. Infiltration rate was recorded using a fairly small rainfall simulator under different vegetation density and terrain mapping units.
The area segmentation was done using aerial photo interpretation, geological map, some satellite images' products and field check.
TMUs show a good association with soil types as well as geological formation, geomorphological origins and the NDVI image.
The sample sites was chosen considering TMUs map, NDVI images and specific site conditions.
The collected infiltration data were shown good fitting using Excell solver based on Horton equation. The estimated initial infiltration for the study area varies between 28-148 mm/hr and constant rate for it varies between 10-50 mm/hr.
Remotely sensed data related to site parameters don't show a high fit because of difficulty of positioning on the images. The date of image recording did not coincide with the date of the fieldwork.
This study demonstrated the complexity of infiltration phenomenon that is governs by different parameters in a mutual interaction.
INTRODUCTION
Infiltration, the passage of water through the upper soil (vadose zone), is an important process in hydrology because it separates rainfall in overland flow and water entering the soil. Overland flow occurs when the rainfall intensity surpasses the infiltration rate. The difference between these two rates is known as the rainfall excess. The overland flow contributes to the direct or fast runoff.
Water, which has entered into the soil either, is held by suction forces or passes through the upper most soil. This causes high suction levels near the surface and thus soil water may move upward.
It is important to know the infiltration because rain fed agriculture depends on the soil moisture. The rainfall excess generates overland flow, and this flow is responsible for soil erosion and transport of soil particles. Effective rainfall is defined as the rainfall, which enters into the soil (i.e. the rainfall minus the direct runoff). Hydrologic modeling requires an insight in the infiltration rates.Infiltration is a complex process. It varies in time and space.
Initial infiltration rates at the start of the rainfall are usually high, if the soil is not saturated or near field capacity .The infiltration rates decline till after some time a fairly constant rate can be observed. Soil suction or sorptivity, as well as effective porosity influence the first part of the curve, i.e. the pores, which allow water to pass under influence of gravity. The proportion and nature of the clays are therefore of importance but also biological activities, which create macropores.
Vegetation is important, vegetation intercept the rainfall and protects soil sealing. Also the biological activity is much higher in vegetated area. The sheltering from radiation and high soil surface temperatures is beneficial to organisms which create macro pores (e.g. canopy density). Infiltration rates are much influenced by the nature of the upper soil, such as sealing, depression storage and stoniness.
Methods and materials:
Study Area:
The selected area for this study is located in the North- East of Tehran (longitude 52°, 15'- 52°, 42' E and latitude 35°, 42'-35°, 57' N). The study area is part of Namrood (Nimroud) a sub catchment of Hableh Rood catchment.
The Namrood catchment area is about 800 km2 and includes 3 main sub catchment, the sub catchment of interest has an area of 78 km2 (7800 ha).
The fan-shaped catchment is in a mountainous area with a maximum elevation of 2740-m and a mean 2370-m amsl elevation, 50% of total area has the elevation above 2000 m.
Figure 1:Study Area Location
The area is surrounded with Siahriz and Sahoun mountains in the North-west and West, Kionchal and Boum mountains in the North, Katouzan, Safidlar, Allahsar, Hazmazgeno and Kharpol mountains in the East, Aabbarik, Chehelcheshmeh and Asbgiran mountains in the South.
Based on the Geological map of Iran (at scale 1:250,000) and field observations, the two major groups of rock recognized in the study area, are sedimentary and volcanic rocks. In the study area geological formations belong to the area from Jurassic to Quaternary. The lithology of the Study Area shows different types including limestone, light gray, bedded to massive partly chert-bearing and dolomitic highly altered diabase, gray bedded limestone, light gray limestone, gray irregular-bedded Sandstone, conglomerate red calcareous, tuff dacitic, green-cream and black shale, talus, scree, rock fall, colluvium, older alluvial fan deposits and young alluvial fan deposits from older to younger.
Figure 2: The Study Area North Direction View
From Geomorphologic point of view five different geomorphologic units occur in the study area including mountains, hills, glacis, plains and valleys .
Considering above mentioned land scapes (geomorphological origins), lithology, local relief and dissection degree the area was categorized to map out TMUs as it shown in figure 3,also a brief classification basics is written on table 1.
According to the small-scale map of "Land Resources and Evaluation map of Tehran ", soil types vary follow the catena. The soils of mountains are mostly shallow to moderately deep with low to high stoniness, whereas the soils of the plains are moderately deep to deep.
Soil texture according to a laboratory test is recorded in 6 different classes, which varies from Clay-loam to Sandy -loam.
Based on Silianof classification method, the area is classified in the slightly semi arid regime that are characterized by a cold winter and mild summer (Lar Co.1993).
The average annual precipitation is about 369 mm based on 20 years collected data at meteorological station, most of it occurs as snow.It is recorded that in the precipitation regime 50% of total precipitation commonly falls during 6 months (December…May, wet and cold months).The absolute minimum and maximum recorded temperature in the Namrood station are respectively -39 c°and +39c°.
Generally natural vegetation components in the south aspects of Alborz Mountains in high elevation consisted of perennial grasses and shrubs. The abundance of annual grasses and forbes vary with fluctuation of yearly precipitation.
Image processing:
Vegetation and its Association with TMUs:
Because of human activities in the study area natural vegetation cover has changed. Rangeland is converted into dry farming and agriculture lands. Crops mainly wheat and barely is grown. There are also some orchards in this area.
The terrain units (TMUs) have been mapped based on aerial photo interpretation.The units have been checked in the field, where geomorphological origins, soil coverage and vegetation data have been observed. The observation sites have been selected based on representative of slope elements.TMU map is shown at figure 3 and it is described by related attribute table1.
Table 1: Geology and Lithology of the Study Area
NDVI map prepration:
The NDVI map of study area is used for analyzing of the spatial variability of infiltration rates considering it's relation with vegetation cover density.
NDVI (Normalized Difference Vegetation Index) values are a measure for the presence and condition of green vegetation.
NDVI values are calculated from two satellites bands, one band containing visible or red reflectance and near-infrared reflectance values.
The NDVI is calculated as follow and NDVI values range from -1 to 1:
NDVI= (near infrared band - visible band)/(near infrared band + visible band)
Vegetated areas will generally yield high values because of their relatively high near-infrared reflectance and low visible reflectance.
In contrast, water, clouds, and snow have larger visible reflectance than near-infrared reflectance. Thus ,these features yield negative index values.
Rock and bare soil areas have similar reflectance in the two bands and result in vegetation indices near zero.
In the time of imagery (early spring) vegetation cover had not grown, then most part of the area, which is used as the rangeland, for local animals grazing, shows low variation of NDVI range. However some part of the area of interest, which is used as cultivated farms or irrigated orchard show high range of NDVI. In the time of imagery most part the area had no much variation of vegetation cover to enhance it in different patches, the red pixels in figure 3 are mostly irrigated farms
Figure 3: NDVI Map of Study Area with TMUs Boundary
Data collection:
Field Data (Infiltration Measurement):
Rainfall simulation at the site, in the units, was done to determine infiltration under different vegetation. Canopy densities have been noted for the small simulation plots.Infiltration rate of different units sampled over the catchment using a simple rainfall simulator apparatus, small thin plexiglas with a volume about 11760 cm3 (28*28*15 cm) and 15 openings (0.5 mm diameter) over a portable metal features. The simulator showers 100 mm/hr as average and it works under variable head, rainfall intensity varies between 60-140 mm/hr during 30 min.
The apparatus is fixed horizontally over the representative plots that are chosen as a representative of landscapes.
A fairly small area of about 1600 cm2 is wetted. Simulation runs lasted for 30 min, some samples for 60 min to see infiltration rate changes at the higher duration of rain shower.
60-field sampling was recorded by 30 (min) simulated shower. Rainfall drops was falling down from 80 cm height, the openings on the bottom of simulator with a diameter equal 0.5 mm made bigger drops than natural rain, had high level of kinetic energy.
Some samples checked on the full vegetation cover and completely bare soil of same point to see how vegetation cover influences water penetration rate trough the soil profile.
Surface runoff and rapid flow through the upper most soil was measured at the down slope part of the plots at various time intervals. Runoff was guided to the collector by a thin metal plate, which is pushed to the soil profile.
A database was built of field sampling results, the data base for each sample was analyzed following association of site characteristics.
Infiltration Data Analysis
Of particular importance is the association between infiltration and canopy density of the vegetation, which was analyzed through scatter plots with textures and slope steepness.
Model Fitting
Horton empirical infiltration model, was used to estimate infiltration rate to elapsed time modified by certain soil properties. Parameters used in the model are estimated from measured infiltration - time relationships for given soil condition. The Horton's model is an equation to estimate infiltration in the following manner:
where i0 and if are the presumed initial and final infiltration rates, and k is an empirical constant and t is time.(EPA,1998)
The recorded runoff data of 30-min sampling was converted to infiltration rate considering simulated rainfall rate. The data set was arranged in the excel files to investigate model fitting on them by solver program.
For each sample three Horton parameter, initial infiltration rate(i0)-constant infiltration rate (if)-and equation exponential ( k )was estimated by excel solver and visual judgment over recorded data graphs.
Horton Equation Parameter Related to the Recorded Site Characteristic Model fitting output will analyze in this stage to investigate how site characteristics follow the infiltration rate of studied units. A comparison will make for initial infiltration rate, constant infiltration rate, equation exponential and vegetation cover, soil particle distribution, and antecedent soil moisture. Also the infiltration observations are related to the NDVI values, to determine the level of accuracy to transform the NDVI image into a map for estimated infiltration.(Figure 4)
RESULTS:
TMU Variations:
Terrain units vary following catena on the different geological units .In the catchment, uplifted fans cover most part of the catchment, which is used for agricultural activities. Some denudational hills include tectonical and eroded valleys. In the some part of area alluvial, colluvial fan and low terraces exist.TMUs interpret geomorphology as well as soil characteristics. The sample points should be located on the proper position same as terrain locations, the marked points on aerial photographs during field work transferred to the images as a point map using different product of images.
The difficulty of positioning for data extraction arose because of NDVI changes in the neighbor pixels, only one pixel shift may report wrong value even while GPS was used.
Vegetation Cover & It's Relation with Detected Indices
Annual grasses in different types cover most part of the area that is grazed by sheep every year early summer. In the time of field observation (October &November) there were not any more except some residual litters. This sort of annual crops grow rapidly every year after springs storm showers. Also some perennial bushes with dark brownish body were found, it has an important effect on infiltration rate over the catchment during showers.
It is clear that recorded vegetation density for each unit doesn't include annual vegetation covers that influence infiltration rate through organic matter gradients of soil and soil aggregations.
The image, which is used in this study, relates to May 1998 and data recording was done in autumn. Also NDVI relations with estimated vegetation cover is not so convenience because of vegetation cover changes during seasons, an it makes lower correlation between NDVI and estimated vegetation for samples points.
An NDVI map made of bands 3&4 TM images (figure 3) have been used to map out coverage condition using remote sensing, the range of NDVI changes (-0.02-+0.20) was not high.
Infiltration Relation with NDVI
Remotely sensed data in the shape of NDVI was related to the different soil categories initial infiltration rate, it is shown in four different groups. Infiltration rate for the same soil textures shows variations in a range between 35 till 140 mm/hr at different locations.
Infiltration Variations
A couple of samples were measured with 60 (min) duration. Most of the samples were measured in a 30 (min) duration to figure out rainfall duration effect on the percolation rate.The infiltration rate curves show a jump in the end part of sampling because of some cracks opening and probably chemical solution of some minerals in the soil body. Infiltration rates shows a good fitting with Horton experimental model(Figure 4) .Initial infiltration rates vary between 28-148 mm/hr in different locations. Estimated constant infiltration rates were recorded in a range between 10-49 mm/hr.
Figure 4: Model Fitting for Sampled Infiltration Data
Field Sampling Observation
Most of the sites have been covered by a thin sealing layer made of small soil particles during previous rainfall happening specially when land slope was moderately gentle with the clay loam soil texture. Water drops caused some microrelief on the soil surface and it hold a part of showered water volume as depression storage.
Pounding stage was not seen at some of the cases, it was because of high range of infiltration on the soil surface but it immediately converted to the sub surface flow and came out of soil profiles as through flow in a depth of less than 5 cm. Bushes and grasses on the surface has shown a good drops trapping, canopy catch rain drops as interception then guide them through the leaves and branches till soil surface, then water drops penetrate to the soil close to the main stem. Water penetration through the soil varied between 5-12 cm in different soil texture and sample situation.
Some of the landscape originated from shale parent material has an appearance with some cracks like crusts on the heavy soil texture after irrigation. These cracks were closing gradually during 30-min sampling. Definitely swelling makes the main role on it to increase infiltration rate. In the some cases loosed soil particles were moved by over land flow to the down slope. There was some deposited sediment on the container that is used for runoff collection.
CONCLUSION and RECOMMENDATIONS
Conclusion
This study was established to survey association between terrestrial evidences and spatial distribution of infiltration in the vadose zone over the catchment . The processing has been implemented based on the integration between field observation data, using a simple rainfall simulator, and remotely sensed images (Landsat TM), using GIS software ILWIS.
Although the field data are shown fit with Horton model it was found that infiltration phenomena through the soil profile shows a complexity that is related to several factors in a mutual interaction to govern infiltration rate.
The level of correlation between studied infiltration parameters and imagery data was not high, it was difficult to position samples locations on the image. Different date of data collection and imagery caused only an envelope curve correlation between terrestrial data, estimated vegetation cover, and satellite recorded data NDVI image.
Sealing effects observed on the sample points has an important effect on recorded infiltration rate specially when the big drops splash soil particles after minutes from sampling start time. The simulator creates water drops with a size of 2.5-3 mm diameter so it is remarkable that splash has an affect on the result.
The spatial variation of infiltration rate in such a mountainous basin like interested site shows much change on the slope to estimate it through the geostatistical approach with 30 m spatial resolution satellite images. Recorded digital number for each pixel of TM images is influenced by different feature in a mixing value.
The imagery season has been early summer so annual grasses effected as the litter on the soil surface. In the images it decreases (NDVI digital number (DN)) in comparison with real grown biomass in the study area.
Soil back ground reflectance, which mostly includes more than 50% pixels' of area, changes reflected DN of bands 3&4 as the original material for NDVI images. It is better to say bands 3&4 differences is low for most of the pixels. Litters and yellowish annual grasses have no effect on band 4 and it is caused some noise in vegetation recorded by imagery especially in the image of interested area in a semi arid condition before first growing of spring.
Recommendations
It is remarkable to consider following comments for the next studies
References:
M.H. Roohian1 , Prof. Dr. A. M.J.Meijerink2 ,Dr.B.Saghafian3 ,P.sajedian4
1 Scientifitic board member of ANRRC of Booshehr province, Iran,
E mail :roohian@yahoo.com
2 Retired head of water resources department of ITC,Enschede,the Netherlands,
E mail:meijerink@itc.nl
3 Scientifitic board member of SCWMRI of Iran,
E mail : saghafian@scwmri
4 The expert of environmental department Fars province office, Iran,
E mail :sajedian@yahoo.com
ABSTRACT
An infiltration experiment was practiced on the collected field data, during autumn 1999, of a semi arid area, Tehran, Iran to indicate level of site parameters associated with measured infiltration rate and remotely sensed data. Infiltration rate was recorded using a fairly small rainfall simulator under different vegetation density and terrain mapping units.
The area segmentation was done using aerial photo interpretation, geological map, some satellite images' products and field check.
TMUs show a good association with soil types as well as geological formation, geomorphological origins and the NDVI image.
The sample sites was chosen considering TMUs map, NDVI images and specific site conditions.
The collected infiltration data were shown good fitting using Excell solver based on Horton equation. The estimated initial infiltration for the study area varies between 28-148 mm/hr and constant rate for it varies between 10-50 mm/hr.
Remotely sensed data related to site parameters don't show a high fit because of difficulty of positioning on the images. The date of image recording did not coincide with the date of the fieldwork.
This study demonstrated the complexity of infiltration phenomenon that is governs by different parameters in a mutual interaction.
INTRODUCTION
Infiltration, the passage of water through the upper soil (vadose zone), is an important process in hydrology because it separates rainfall in overland flow and water entering the soil. Overland flow occurs when the rainfall intensity surpasses the infiltration rate. The difference between these two rates is known as the rainfall excess. The overland flow contributes to the direct or fast runoff.
Water, which has entered into the soil either, is held by suction forces or passes through the upper most soil. This causes high suction levels near the surface and thus soil water may move upward.
It is important to know the infiltration because rain fed agriculture depends on the soil moisture. The rainfall excess generates overland flow, and this flow is responsible for soil erosion and transport of soil particles. Effective rainfall is defined as the rainfall, which enters into the soil (i.e. the rainfall minus the direct runoff). Hydrologic modeling requires an insight in the infiltration rates.Infiltration is a complex process. It varies in time and space.
Initial infiltration rates at the start of the rainfall are usually high, if the soil is not saturated or near field capacity .The infiltration rates decline till after some time a fairly constant rate can be observed. Soil suction or sorptivity, as well as effective porosity influence the first part of the curve, i.e. the pores, which allow water to pass under influence of gravity. The proportion and nature of the clays are therefore of importance but also biological activities, which create macropores.
Vegetation is important, vegetation intercept the rainfall and protects soil sealing. Also the biological activity is much higher in vegetated area. The sheltering from radiation and high soil surface temperatures is beneficial to organisms which create macro pores (e.g. canopy density). Infiltration rates are much influenced by the nature of the upper soil, such as sealing, depression storage and stoniness.
Methods and materials:
Study Area:
The selected area for this study is located in the North- East of Tehran (longitude 52°, 15'- 52°, 42' E and latitude 35°, 42'-35°, 57' N). The study area is part of Namrood (Nimroud) a sub catchment of Hableh Rood catchment.
The Namrood catchment area is about 800 km2 and includes 3 main sub catchment, the sub catchment of interest has an area of 78 km2 (7800 ha).
The fan-shaped catchment is in a mountainous area with a maximum elevation of 2740-m and a mean 2370-m amsl elevation, 50% of total area has the elevation above 2000 m.
Figure 1:Study Area Location
The area is surrounded with Siahriz and Sahoun mountains in the North-west and West, Kionchal and Boum mountains in the North, Katouzan, Safidlar, Allahsar, Hazmazgeno and Kharpol mountains in the East, Aabbarik, Chehelcheshmeh and Asbgiran mountains in the South.
Based on the Geological map of Iran (at scale 1:250,000) and field observations, the two major groups of rock recognized in the study area, are sedimentary and volcanic rocks. In the study area geological formations belong to the area from Jurassic to Quaternary. The lithology of the Study Area shows different types including limestone, light gray, bedded to massive partly chert-bearing and dolomitic highly altered diabase, gray bedded limestone, light gray limestone, gray irregular-bedded Sandstone, conglomerate red calcareous, tuff dacitic, green-cream and black shale, talus, scree, rock fall, colluvium, older alluvial fan deposits and young alluvial fan deposits from older to younger.
Figure 2: The Study Area North Direction View
From Geomorphologic point of view five different geomorphologic units occur in the study area including mountains, hills, glacis, plains and valleys .
Considering above mentioned land scapes (geomorphological origins), lithology, local relief and dissection degree the area was categorized to map out TMUs as it shown in figure 3,also a brief classification basics is written on table 1.
According to the small-scale map of "Land Resources and Evaluation map of Tehran ", soil types vary follow the catena. The soils of mountains are mostly shallow to moderately deep with low to high stoniness, whereas the soils of the plains are moderately deep to deep.
Soil texture according to a laboratory test is recorded in 6 different classes, which varies from Clay-loam to Sandy -loam.
Based on Silianof classification method, the area is classified in the slightly semi arid regime that are characterized by a cold winter and mild summer (Lar Co.1993).
The average annual precipitation is about 369 mm based on 20 years collected data at meteorological station, most of it occurs as snow.It is recorded that in the precipitation regime 50% of total precipitation commonly falls during 6 months (December…May, wet and cold months).The absolute minimum and maximum recorded temperature in the Namrood station are respectively -39 c°and +39c°.
Generally natural vegetation components in the south aspects of Alborz Mountains in high elevation consisted of perennial grasses and shrubs. The abundance of annual grasses and forbes vary with fluctuation of yearly precipitation.
Image processing:
Vegetation and its Association with TMUs:
Because of human activities in the study area natural vegetation cover has changed. Rangeland is converted into dry farming and agriculture lands. Crops mainly wheat and barely is grown. There are also some orchards in this area.
The terrain units (TMUs) have been mapped based on aerial photo interpretation.The units have been checked in the field, where geomorphological origins, soil coverage and vegetation data have been observed. The observation sites have been selected based on representative of slope elements.TMU map is shown at figure 3 and it is described by related attribute table1.
NDVI map prepration:
The NDVI map of study area is used for analyzing of the spatial variability of infiltration rates considering it's relation with vegetation cover density.
NDVI (Normalized Difference Vegetation Index) values are a measure for the presence and condition of green vegetation.
NDVI values are calculated from two satellites bands, one band containing visible or red reflectance and near-infrared reflectance values.
The NDVI is calculated as follow and NDVI values range from -1 to 1:
NDVI= (near infrared band - visible band)/(near infrared band + visible band)
Vegetated areas will generally yield high values because of their relatively high near-infrared reflectance and low visible reflectance.
In contrast, water, clouds, and snow have larger visible reflectance than near-infrared reflectance. Thus ,these features yield negative index values.
Rock and bare soil areas have similar reflectance in the two bands and result in vegetation indices near zero.
In the time of imagery (early spring) vegetation cover had not grown, then most part of the area, which is used as the rangeland, for local animals grazing, shows low variation of NDVI range. However some part of the area of interest, which is used as cultivated farms or irrigated orchard show high range of NDVI. In the time of imagery most part the area had no much variation of vegetation cover to enhance it in different patches, the red pixels in figure 3 are mostly irrigated farms
Figure 3: NDVI Map of Study Area with TMUs Boundary
Data collection:
Field Data (Infiltration Measurement):
Rainfall simulation at the site, in the units, was done to determine infiltration under different vegetation. Canopy densities have been noted for the small simulation plots.Infiltration rate of different units sampled over the catchment using a simple rainfall simulator apparatus, small thin plexiglas with a volume about 11760 cm3 (28*28*15 cm) and 15 openings (0.5 mm diameter) over a portable metal features. The simulator showers 100 mm/hr as average and it works under variable head, rainfall intensity varies between 60-140 mm/hr during 30 min.
The apparatus is fixed horizontally over the representative plots that are chosen as a representative of landscapes.
A fairly small area of about 1600 cm2 is wetted. Simulation runs lasted for 30 min, some samples for 60 min to see infiltration rate changes at the higher duration of rain shower.
60-field sampling was recorded by 30 (min) simulated shower. Rainfall drops was falling down from 80 cm height, the openings on the bottom of simulator with a diameter equal 0.5 mm made bigger drops than natural rain, had high level of kinetic energy.
Some samples checked on the full vegetation cover and completely bare soil of same point to see how vegetation cover influences water penetration rate trough the soil profile.
Surface runoff and rapid flow through the upper most soil was measured at the down slope part of the plots at various time intervals. Runoff was guided to the collector by a thin metal plate, which is pushed to the soil profile.
A database was built of field sampling results, the data base for each sample was analyzed following association of site characteristics.
Infiltration Data Analysis
Of particular importance is the association between infiltration and canopy density of the vegetation, which was analyzed through scatter plots with textures and slope steepness.
Model Fitting
Horton empirical infiltration model, was used to estimate infiltration rate to elapsed time modified by certain soil properties. Parameters used in the model are estimated from measured infiltration - time relationships for given soil condition. The Horton's model is an equation to estimate infiltration in the following manner:
where i0 and if are the presumed initial and final infiltration rates, and k is an empirical constant and t is time.(EPA,1998)
The recorded runoff data of 30-min sampling was converted to infiltration rate considering simulated rainfall rate. The data set was arranged in the excel files to investigate model fitting on them by solver program.
For each sample three Horton parameter, initial infiltration rate(i0)-constant infiltration rate (if)-and equation exponential ( k )was estimated by excel solver and visual judgment over recorded data graphs.
Horton Equation Parameter Related to the Recorded Site Characteristic Model fitting output will analyze in this stage to investigate how site characteristics follow the infiltration rate of studied units. A comparison will make for initial infiltration rate, constant infiltration rate, equation exponential and vegetation cover, soil particle distribution, and antecedent soil moisture. Also the infiltration observations are related to the NDVI values, to determine the level of accuracy to transform the NDVI image into a map for estimated infiltration.(Figure 4)
RESULTS:
TMU Variations:
Terrain units vary following catena on the different geological units .In the catchment, uplifted fans cover most part of the catchment, which is used for agricultural activities. Some denudational hills include tectonical and eroded valleys. In the some part of area alluvial, colluvial fan and low terraces exist.TMUs interpret geomorphology as well as soil characteristics. The sample points should be located on the proper position same as terrain locations, the marked points on aerial photographs during field work transferred to the images as a point map using different product of images.
The difficulty of positioning for data extraction arose because of NDVI changes in the neighbor pixels, only one pixel shift may report wrong value even while GPS was used.
Vegetation Cover & It's Relation with Detected Indices
Annual grasses in different types cover most part of the area that is grazed by sheep every year early summer. In the time of field observation (October &November) there were not any more except some residual litters. This sort of annual crops grow rapidly every year after springs storm showers. Also some perennial bushes with dark brownish body were found, it has an important effect on infiltration rate over the catchment during showers.
It is clear that recorded vegetation density for each unit doesn't include annual vegetation covers that influence infiltration rate through organic matter gradients of soil and soil aggregations.
The image, which is used in this study, relates to May 1998 and data recording was done in autumn. Also NDVI relations with estimated vegetation cover is not so convenience because of vegetation cover changes during seasons, an it makes lower correlation between NDVI and estimated vegetation for samples points.
An NDVI map made of bands 3&4 TM images (figure 3) have been used to map out coverage condition using remote sensing, the range of NDVI changes (-0.02-+0.20) was not high.
Infiltration Relation with NDVI
Remotely sensed data in the shape of NDVI was related to the different soil categories initial infiltration rate, it is shown in four different groups. Infiltration rate for the same soil textures shows variations in a range between 35 till 140 mm/hr at different locations.
Infiltration Variations
A couple of samples were measured with 60 (min) duration. Most of the samples were measured in a 30 (min) duration to figure out rainfall duration effect on the percolation rate.The infiltration rate curves show a jump in the end part of sampling because of some cracks opening and probably chemical solution of some minerals in the soil body. Infiltration rates shows a good fitting with Horton experimental model(Figure 4) .Initial infiltration rates vary between 28-148 mm/hr in different locations. Estimated constant infiltration rates were recorded in a range between 10-49 mm/hr.
Figure 4: Model Fitting for Sampled Infiltration Data
Field Sampling Observation
Most of the sites have been covered by a thin sealing layer made of small soil particles during previous rainfall happening specially when land slope was moderately gentle with the clay loam soil texture. Water drops caused some microrelief on the soil surface and it hold a part of showered water volume as depression storage.
Pounding stage was not seen at some of the cases, it was because of high range of infiltration on the soil surface but it immediately converted to the sub surface flow and came out of soil profiles as through flow in a depth of less than 5 cm. Bushes and grasses on the surface has shown a good drops trapping, canopy catch rain drops as interception then guide them through the leaves and branches till soil surface, then water drops penetrate to the soil close to the main stem. Water penetration through the soil varied between 5-12 cm in different soil texture and sample situation.
Some of the landscape originated from shale parent material has an appearance with some cracks like crusts on the heavy soil texture after irrigation. These cracks were closing gradually during 30-min sampling. Definitely swelling makes the main role on it to increase infiltration rate. In the some cases loosed soil particles were moved by over land flow to the down slope. There was some deposited sediment on the container that is used for runoff collection.
CONCLUSION and RECOMMENDATIONS
Conclusion
This study was established to survey association between terrestrial evidences and spatial distribution of infiltration in the vadose zone over the catchment . The processing has been implemented based on the integration between field observation data, using a simple rainfall simulator, and remotely sensed images (Landsat TM), using GIS software ILWIS.
Although the field data are shown fit with Horton model it was found that infiltration phenomena through the soil profile shows a complexity that is related to several factors in a mutual interaction to govern infiltration rate.
The level of correlation between studied infiltration parameters and imagery data was not high, it was difficult to position samples locations on the image. Different date of data collection and imagery caused only an envelope curve correlation between terrestrial data, estimated vegetation cover, and satellite recorded data NDVI image.
Sealing effects observed on the sample points has an important effect on recorded infiltration rate specially when the big drops splash soil particles after minutes from sampling start time. The simulator creates water drops with a size of 2.5-3 mm diameter so it is remarkable that splash has an affect on the result.
The spatial variation of infiltration rate in such a mountainous basin like interested site shows much change on the slope to estimate it through the geostatistical approach with 30 m spatial resolution satellite images. Recorded digital number for each pixel of TM images is influenced by different feature in a mixing value.
The imagery season has been early summer so annual grasses effected as the litter on the soil surface. In the images it decreases (NDVI digital number (DN)) in comparison with real grown biomass in the study area.
Soil back ground reflectance, which mostly includes more than 50% pixels' of area, changes reflected DN of bands 3&4 as the original material for NDVI images. It is better to say bands 3&4 differences is low for most of the pixels. Litters and yellowish annual grasses have no effect on band 4 and it is caused some noise in vegetation recorded by imagery especially in the image of interested area in a semi arid condition before first growing of spring.
Recommendations
It is remarkable to consider following comments for the next studies
- Date of imagery and data collection duration for dry lands should be in a season and moisture condition.
- It is perfect to record field data of studies related to canopy density in the season with maximum vegetation cover.
- Sample point position should be record by an accurate GPS for positioning points on the images.
- Rainfall simulator construction would be resulted more reliable if local condition were considered.
References:
- America society of civil engineers,1996.Hydrology Handbook.Report No 28
- Applied Meteorological Research Center, 1995, Intensity-Duration- Frequency relations at synoptic stations in Iran.
- Bergsma,E.P. Charman, F.Gibbons, H.Hurni, W.C.Moldenhaver, S.Panichapong, 1996. Terminology for soil erosion and conservation. ITC. publications, Enschede. The Netherlands.
- Casenave, A. and C.Valentin,1992.Arunoff capability clasification system based on a surface features criteria in semi arid areas of West Africa.J.Hydrol.130,231-249.
- Fox D.M., Y.Le Bissonnais, A.Bruand,1998.The effect of ponding depth on infiltration in a crusted surface depression.Catena.32:87-100.
- Hillel, Daniel 1999.Environmental soil physics.401
- Kamphorst A.,1987.A small rainfall simulator for the determination of soil readability. Netherlands Journal of Agricultural Siences.35,407-415.
- Sharma M.L, Gander,G.A, and Hunt, C.G., 1980.Spatial variability of infiltration in a watershed. J.Hydrol., 45: 101-122.
- Meijerink, Allard M.J., H. A. M. de Brouwer, C.M.Mannaerts., 1994. Introduction to the use of GIS for practical hydrology, 49-147.
- Schiffler, Gerd R. 1990. Conceot for the investigation of the spatial variability of infiltration parameters in a small catchment. International workshop on geostatistical methods: recent developments and application in surface and subsurface hydrology.
- Sullivan, M. J. J. warwick, S.W. Tyler,1996. Quantity and delineating spatial variations of surface infiltration in a small watershed. J.Hydrol.,181:149-168.
- Viessmann, Warren and Knapp, Jr, J.W. , Gary L. Lewis and Terence E. Harbaugh.1977. Introduction to hydrology.
- Williams, Joseph R. Ying Ouyang and Jin-Song Chen,1998.Estimation of infiltration rate in the vadose zone: Application of selected mathematical models. Vol I&II:1-10.
- WOOD Eric F. et.al, 1987.Effect of spatial variability and scale with implications to hydrologic modeling. J. Hydrol., 102:29-47.