Monitoring and modelling of irrigation
practices and its effect on catchment hydrology
for fragmented landscapes
An application example in the Ethiopian Rift Valley
Master Thesis
Eva Hoekstra
September 2016 – June 2017
Author: Eva Hoekstra
Student number: 3771385
First supervisor: Marjolein Vogels
Second supervisor: Elisabeth Addink
Earth Surface and Water
Geohazards and Earth Observation
Faculty of Geosciences
Utrecht University
Abstract
It is known that irrigation can significantly alter the distribution of water in a terrestrial system. Soit is of importance to include the presence of irrigation in hydrological modelling studies. However,in many countries, consistent and reliable documentation on artificial water applications is notavailable. Previous studies successfully mapped irrigation with remote sensing by using a pixel-based approach on NDVI information. Object-based image analysis has often proved to producemore accurate results, especially when working with high resolution imagery in spatially andspectrally complex areas. Therefore, an object-based remote sensing method to identify irrigatedobjects in a fragmented landscape, located in the Ethiopian Rift Valley, was developed and tested.SPOT imagery with a resolution of 1.5 meter was used and resulted in a highly satisfying irrigationmap. The same irrigation mapping method was performed on coarser resolution imagery (Landsat).Evidently, SPOT was able to capture the small fragmented agricultural areas more accurately thanLandsat. However, despite the coarser resolution of Landsat imagery, a visual inspection of theSPOT and Landsat irrigation maps revealed that Landsat also had its advantages. It performedbetter in the classification of heterogeneous agricultural fields and bare agricultural plots. Theobject-based irrigation mapping method developed in this study has proven to be an adequateapproach for identifying irrigated agriculture, with the use of both 1.5 and 30-meter resolutionimagery. It enables the production of irrigation maps with a relatively high temporal frequency,causing it to be a promising input for hydrological modelling studies. Remotely sensed irrigationmaps were implemented in the Spatial Processes in HYdrology (SPHY) model to examine the impactof irrigation on the outcomes of the model. Two different scenarios were set up for a case study inthe Tikur Woha catchment in Ethiopia. The reference scenario did not consider the application ofirrigation while the irrigation scenario incorporated spatio-temporal irrigation information.Increases in soil moisture content, evapotranspiration, root percolation, and to a lesser extentrunoff, were observed in the catchment when the presence of irrigation was taken into account. Theremotely sensed irrigation maps produced in this study indicated that the extent of irrigatedagriculture is highly variable over time. So the modelling results emphasized that when irrigation isused as input for hydrological models it must be properly characterized, both spatially andtemporally, as it can contribute considerably to the uncertainty of model outcomes.
Keywords Irrigation · Fragmented agriculture · Remote sensing · OBIA · Hydrological modelling · SPHY
Table of contents
Acknowledgements........................................................................................ ii
Abstract .............................................................................................................iii
List of tables.................................................................................vii
List of figures ..............................................................................................vii
1. Introduction................................................................................................... 1
1.1 Mapping irrigation with remote sensing .......................1
1.1.1 Definition and aspects of irrigation ............................................2
1.1.2 Spatial resolution requirements..........................................................................................................2
1.2 Incorporation of irrigation in hydrological modelling......................................3
1.2.1 Influence of irrigation on the water balance........................................................3
1.2.2 Implementing irrigation information in hydrological models........................................4
1.3 Study area: Ethiopia..................................................................................................5
1.4 Knowledge gap...................................................................................................5
1.5 Research aim and objectives......................................................................................6
2. Site description........................................................................................ 7
2.1 Geographical setting........................................................................................7
2.1.1 Lake Basaka area .................................................................................................7
2.1.2 Tikur Woha catchment.......................................................................7
2.2 Topography and soil.........................................................................9
2.3 Climate .........................................................................................9
2.4 Land use............................................................................................. 10
2.4.1 Land use around Lake Basaka.......................................................... 11
2.4.2 Land use in the Tikur Woha catchment...................................... 11
3. Methodology..............................................................................12
3.1 Data collection........................................................................................... 12
3.1.1 SPOT data.................................................................................................... 12
3.1.2 Landsat data ......................................................................................................... 12
3.1.3 Meteorological data .................................................................................. 13
3.1.4 Topographical and soil data.................................................................... 13
3.1.5 Land-cover data ....................................................................................... 13
3.2 Image pre-processing................................................................................. 14
v
3.3 Image segmentation .............................................................. 14
3.3.1 Parameter settings for the segmentation.............................................. 15
3.4 Classification in the Lake Basaka area ................................................... 16
3.4.1 Training- and validation data ................................................ 18
3.4.2 Random Forest.................................................................... 19
3.5 Classification in the Tikur Woha catchment........................................................................................... 19
3.6 Comparison SPOT and Landsat irrigation maps in the Lake Basaka area ............ 20
3.7 Hydrological modelling with SPHY ................................................. 20
3.7.1 The SPHY model.............................................................................. 20
3.7.2 Translating spatio-temporal irrigation data into input for SPHY ..................... 21
3.7.3 Implementation of irrigation within SPHY......................................................... 23
3.7.4 Comparison of the reference and irrigation scenarios................................. 24
4. Results..............................................................................25
4.1 Irrigation map based on SPOT imagery for the Lake Basaka area ..................... 25
4.2 Irrigation map based on Landsat imagery for the Lake Basaka area........................ 27
4.3 Comparison of the SPOT and Landsat irrigation maps in the Lake Basaka area............. 29
4.3.1 Large agricultural fields........................................................ 32
4.3.2 Small agricultural fields.......................................................................... 35
4.3.3 Variable importance during the classification ........................ 37
4.4 Irrigation maps in the Tikur Woha catchment..................................... 38
4.5 Simulated irrigation applications and abstractions in the Tikur Woha catchment............. 39
4.6 Irrigation impact on the hydrology of the Tikur Woha catchment.............................. 41
4.6.1 Changes in soil moisture content due to irrigation .................. 42
4.6.2 Changes in evapotranspiration due to irrigation.......................... 44
4.6.3 Changes in root percolation due to irrigation ........................... 46
4.6.4 Changes in runoff due to irrigation.................................. 47
5. Discussion..................................................................50
5.1 Significance of monitoring and modelling irrigation practices ......................... 50
5.2 Mapping performance of a SPOT-based and Landsat-based irrigation map ............... 50
5.2.1 Delineation quality of the segmentation ............................. 50
5.2.2 Classification accuracy and uncertainties.............................. 51
5.2.3 Importance of variables in the classification..................................... 51
5.2.4 Differences between the SPOT and Landsat irrigation map ......................... 51
vi
5.3 Effect spatio-temporal irrigation information on hydrological modelling.................... 52
5.4 Recommendations and future study areas .................... 53
5.4.1 Recommendations for modelling irrigated agriculture ........................... 53
5.4.2 Future interesting study areas........................................... 53
6. Conclusion................................................................55
References ..................................................................56
Appendices ........................................................................62
1A: Variable importance plots of the Landsat classification........................... 62
1B: Variable importance plots of the SPOT classification ..................................... 63
1C: Daily irrigation applications as simulated by the irrigation scenario in SPHY............ 64
Figure 1.1: Illustration of the visual difference between irrigated and non-irrigated areas. Image A: centre-pivot
irrigation in the oasis region Al Ahsa in eastern Saudi Arabia. The irrigated crop circles clearly contrast with the
surrounding area. Image B: small-scale fragmented agriculture in Ethiopia, located south of Lake Basaka. The
contrast between irrigated and unirrigated fields is less extreme but still visible.
Figure 2.1: (A) Location of the two study areas within Ethiopia, (B) the study area around Lake Basaka, and (C)
the Tikur Woha catchment.
Figure 2.2: Digital elevation model of (A) the Lake Basaka area and (B) the Tikur Woha catchment.
Figure 2.3: Soil map of the Tikur Woha catchment. The legend also includes the appearance of the different soil
types in percentages.
Figure 4.1: Final irrigation map created with SPOT imagery.
Figure 4.2: Final irrigation map created with Landsat imagery.
Figure 4.3: Correspondence of Landsat objects with the SPOT classification. The percentage classes indicate for
each Landsat image object the agreement with the SPOT irrigation map.
Full Text
Download Click here
ليست هناك تعليقات:
إرسال تعليق