Effect of Microtopography on Rill Incision at Field Scale Using Close Range Photogrammetry at Angereb Watershed in Ethiopia G.D. Gessesse (1), A. Klik (1), H. Fuchs (2), R. Mansberger (2), and D. Rieke-Zapp (3) (1) Institute of Hydraulics and Rural Water Management.
[email protected] , (2) Institute of Surveying, Remote Sensing, and Land Information, University of Natural Resources and Applied Life Sciences, Vienna, Austria., (3) Institute of Geological Sciences, University of Bern, Switzerland 0.30
0.40
Introduction
Universität für Bodenkultur Wien Department für W asser-AtmosphäreUmwelt Institut für Hydraulik und landeskulturelle W asserwirtschaft
Clay loam S lo p e g ra d ie n t ( m m
The difficulty in measuring hydraulic erosion processes during runoff events at field condition necessitates the use of spatial topographic and micro-topographic factors to characterize rill initiation. Yet, the spatial rill formation and development are poorly understood at field condition where these factors are more complex, and occur at different intensities. The objective was to quantify the compound effect of topography and micro-topography on rill incision.
S lo p e g ra d ie n t (m m
-1
-1
)
)
Sandy loam 0.30
0.20 0.264
S = 0.6211*RR 2
0.25
0.20 0.346
S= 0.8261*RR R2 = 0.5524
0.15
R = 0.3993 0.10 0.01
0.02
0.03
0.04
0.05
0.10 0.06 0.01
0.02
0.03
RR (m)
0.04
0.05
RR (m)
Fig. 1. Relation between slope gradient and random roughness
Materials and methods • The experiment was conducted on freshly tilled, bare plots (3.5 m by 13.5 m) with sandy loam and clay loam textures in Angereb watershed north of lake Tana, Ethiopia in July 2007.
0.80
b
S=a*Ac *RR
c
0.74
a
b
0.60 0.55 0.40 0.20
0.31
0.26
0.00 Sandy loam (random roughness)
Clay loam (oriented roughness)
-0.20
-0.03
-0.08 Sandy loam
c Fig. 2. Rill initiation parameters: a=threshold drainage area, b= relative area exponent, c= relative roughness exponent
Clay loam
• DEMs of 1 cm grid size were derived from close range photogrammetry using a calibrated Canon EOS 1Ds digital single lens reflex camera.
• DEMs were generated before and after 42 mm (92 Jm-2 of energy) rainfall. • 25-35 rills with depth 3 - 8 cm and contributing areas of 0.4-4 m2 were used for rill initiation analysis. • Unit rill contributing area (Ac, m2 m-1), mean slope gradient (Scr, m m-1) and random roughness before the rainfall (RR, m) at head cut of rills, and rill network were derived from the DEM. • Rill incision parameters were estimated from non-linear regression, Scr = a*Acb*RRc. • Critical rill areas were identified where Scr*Ac-b*RR-c exceeds the constant, ‘a’.
Results and discussion Critical rill erosion (a minimum of 1kg soil loss) occurred where Scr*Ac-b*RR-c exceeds 0.55 and 0.74 for sandy loam and clay loam textures. Topographic surface area required to initiate critical rill erosion is greater for clay loam than sandy loam texture. The predicted relative area exponent is significantly different from the relative area exponent (-0.40) for rill erosion reported at field study in Belgium (Govers, 1991), indicating the scale dependent of the parameters. The non-linear relationship was highly explained by soil surface roughness than the unit contributing area High rill density and rill initiation occurred in areas where the roughness was high (Fig.3)
Sandy loam
Index 0.55
Index > 0.74 Clay loam
Fig. 3. Spatial distribution of soil surface roughness and digitized rill network and critical rill erosion source areas (white shaded areas)
Conclusion The results indicate the merits in the consideration of micro-topography in delineating areas prone to further rilling at micro-scale using high resolution DEMs. Threshold drainage area for rill initiation and the rill network for a given slope are related to the size, orientation and distribution of the surface roughness. Extrapolation of the result is limited because of the scale effect of the parameters (the threshold drainage area is more sensitive to the change in slope gradient).
Acknowledgment The study was financially supported by Austrian Development Cooperation. The IVFL of BOKU, Vienna has provided the full support of the photogrammetry analysis. Special thanks to Dr. Zapp for providing the camera.
