effect of vegetative cover on usda curve numbers for ... · in the australian semi-arid tropics j....
TRANSCRIPT
DEPARTMENT OF LANDS, PLANNING AND ENVIRONMENT
Technical Memorandum
EFFECT OF VEGETATIVE COVER ON USDA CURVE NUMBERS FOR PASTORAL CATCHMENTS
IN THE AUSTRALIAN SEMI-ARID TROPICS
J. A. Motha and M. Dilshad
Land Resource Conservation Branch Darwin, Australia
DLPE Technical Memorandum Number 97/01
ISBN 0 7245 2976 4
June 1997
EFFECT OF VEGETATIVE COVER ON USDA CURVE NUMBERS FOR PASTORAL CATCHMENTS
IN THE AUSTRALIAN SEMI-ARID TROPICS
J. A. Motha and M. Dilshad
Keywords: Runoff, Curve number, Cover, GRASP, LAMSA T, Semi-arid tropics, Erosion, Land degradation, Northern Territory
Published by: Land Resource Conservation Branch, Land Resources Division Department of Lands, Planning and Environment, GPO Box 1680, Darwin, NT 0801, Australia.
Abstract
The Australian semi-arid tropics (SAT) expenences highly eroslve rainstorms
possessing the potential to cause much damage to poorly managed lands. The
LAMSAT model, developed to assess production and associated land degradation in
the SAT, was used to study the relationship between cover and runoff for pastoral
catchments in the Northern Territory. A good treatment of the statistical indicators for
the optimisation of the runoff parameter CN2 is presented. The scatter in the cover
CN2 function obtained is believed to be due to the differences in the cultural history of
the catchments studied. Useful information concerning studies similar to this study
have also been captured in this report. Directions for possible further work on this
theme have also been identified.
Contents
1. Introduction ..... ... .. .. ... .... ............ .. .... .... ... .......... .... .. ....... .. ... .... ....... .... ......... ............ l
2. The LAMSAT Runoff Sub-Model ...... .. .. .......... ........ ............ .. .. .... .... ...... .. ............ 2
3. The Cover-CN2 Function ..... ... ..... .. .. ..... ... .. .. ... ... .. ... ... .. ....... .... ........... .... .... ..... ... ... 2
3.1. Cover-CN2 Function for Data Collected at Douglas Daly .... .. ........... .. )
3.1.1. Using Runoff Data over Segments of a Season .. .. .... ... ..... .. .. .. .. . )
3.1.2. Using Runoff Data over a Season with Manipulated Cover .... .. 5
3.2. Other Cover-CN2 Functions .... .... ..... .. ..... .... .... ............ .... ..... .. ..... ..... .. ..... . 8
3.3 . Sensitivity of Frequency of Runoff Comparison on the Cover-CN2
Function .... ... ........ ... .... .... ...... ....... .. .. .... ...... ... ... ... ........ .. ...... ............ ....... .... 9
4. Discussion ... ... ... .. .... .... ... .......... .. .......... .. ......... ... .... .. .... ... .. ...... ....... ...... .... .. ..... ... .... . 10
5. Conclusions & Recommendations ... ... ....... .. ...... .. ..... .. .......... .. .. ..... ... .... ...... ....... .... 12
6. Acknowledgments ...... ... .. ... ......... ........ ..... .. ..... .. ... ... .... .... .... .... ..... ... ...... .... ..... ... ... .. 12
7. References ..... .... ... ... ........ ....... .... ....... ... .. ... .. ..... ...... ... ...... ...... ....... .. .... ...... ......... .. ... . 13
Effect of vegetative cover on USDA curve numbers
1. Introduction
Pastoral industry is the predominant agricultural activity in the Australian semi-arid
tropics (SAT). Pastoral lands in the SAT are, however, potentially susceptible to serious
land degradation, especially soil erosion. This is because the region is characterised by high
energy rainfalls and strong variability and seasonality in climate and plant growth
responses.
To facilitate sustainable utilisation of the land resources of the SAT, the temporal and
spatial dynamics of the interactions between soil, water and plant need to be understood.
Furthermore, mechanisms that allow the prediction of the impact of various land conditions
and management strategies on production and degradation are required. Computer
simulation models provide powerful tools to assess spatial and temporal dynamics of
complex interactive relationships as well as to predict the impacts of various
influencing factors on the risks and magnitude of production and associated
environmental degradation.
The LAMSAT model (Dilshad et aI., in preparation) was developed for predicting,
within the context of seasonal as well as long term variabilities and trends, the impacts
of land management strategies on beef cattle production, water fluxes and soil erosion
in the SAT. The LAMSAT model consists of a pasture growth sub-model, an erosion
sub-model and a runoff sub-model.
The runoff sub-model in LAMSAT is a modified version of the Curve Number Method
(Williams and LaSeur, 1976) developed by the Soil Conservation Service of the United
States Department of Agriculture (USDA-SCS). Two important parameters required
by the LAMSAT runoff sub-model are, the curve number for average soil moisture
condition (CN2) for bare soil (CN2bare) and the slope of a linear relationship between
CN2 and cover, which together define the cover-CN2 function (Yee Yet, 1994).
This report briefly presents the procedure involved in obtaining the cover-CN2
function using the LAMSAT model and discusses the effect of cover on CN2.
1
Effect of vegetative cover on USDA curve numbers
2. The LAMSAT Runoff Sub-Model
The Curve Number Method (CNM) is described in a number of publications
(Boughton, 1989; Littleboy et aI. , 1989; Dilshad and Peel, 1994; and Yee Yet, 1994).
It has been used extensively to predict surface runoff on agricultural catchments in a
wide range of climatic regions (Boughton, 1989). The method is simple,
computationally efficient and requires inputs that are generally available. Hence, a
number of agricultural simulation models have incorporated the CNM in its various
forms (Dilshad and Peel, 1994). Despite its wide use, an inherent limitation of the
CNM is that it is essentially empirical in nature and lacks a sound physical basis
(Boughton, 1989). Another major limitation of the method was that the dynamic effect
of cover on runoff was not accounted for. Littleboy et ai. (1989) and Y ee Yet (1994),
working within different modelling frameworks, incorporated this aspect into the
method. Whilst Littleboy et ai. (1989) dealt with croplands, Yee Yet (1994) analysed
pastoral lands. The modified version of the CNM reported by Yee Yet (1994) has been
adopted for the LAMSAT runoff sub-model.
The original CNM (Williams and LaSeur, 1987) required CN2 as an input and
appropriately modified its value to account for the moisture conditions prevalent on a
given day. In order to address the dynamic nature of surface cover in agricultural
catchments, CN2 for zero cover (CN2bare) and the slope of a cover-CN2 relationship
were made inputs in the modified version (Yee Yet, 1994). Provision was also made
for CN2 to be a constant beyond a specified cover. With these enhancements, the sub
model obtains a CN2 value corresponding to the cover on a given day and then
accounts for the soil moisture, to arrive at the value for daily runoff.
In the ensuing section, the cover-CN2 relationship for pastoral catchments on
Kandosols at the Douglas Daly Research Farm (latitude 13°5 I's, longitude 131°1z'E,
elevation approximately 50 m above sea level) is derived.
3. The Cover-CN2 Function
To derive the cover-CN2 relationship, optimised CN2 values are required at different,
relatively constant covers. These points are then plotted to obtain the function in the
2
Effect of vegetative cover on USDA curve numbers
prescribed form - linear decrease of CN2, from its value at zero or near-zero cover, up
to a certain high cover beyond which CN2 remains a constant.
The procedure for optimising CN2 involves using the LAMSAT model with the slope
of the cover-CN2 function equated to zero. This implies that CN2 remains constant for
the period of simulation. By trial and error, the optimal value of CN2 was obtained for
the best agreement between predicted and observed runoff and soil moisture whilst
making sure that the predicted and observed cover and yield too were in good
agreement. The criteria for optimising the CN2 were to minimise root mean square
error (RMSE) and average error, while simultaneously examining correlation between
individual predicted and observed values and the ratio of total predicted to total
observed value. Another important issue in the optimisation of CN2 was that the
upper limit to soil evaporation EPLIl\.1 (in the pasture growth sub-model) affects the
predicted runoff, and in tum, the optimised CN2. Therefore, it is advisable to pivot
EPLIM at a particular value whilst the cover-CN2 function is being derived for a
scenario of interest and that this EPLIl\.1 value be associated with that cover-CN2
function, for future use on similar scenarios.
3.1. Cover-CN2 Function for Data Collected at Douglas Daly
3.1.1. Using Runoff Data over Segments of a Season
Initially data from pasture growth trials conducted at the Douglas Daly Research Farm
(peel et aI. , 1995) were used to obtain the cover-CN2 function. These data were
collected from two 20 m x 20 m areas, fenced to protect from cattle, within each of
three catchments (5 .9, 6.8, 7.8 ha in size) under improved pasture (Urochloa
mosambicensis) . Vegetation cover, yield, soil moisture and plant nutrient data were
collected from September 1992 to August 1993. The data collection was phased in
such a way to cover different regimes of the pasture growth. As the variation between
trial plots within each catchment was found to be insignificant, the mean of the data
from the two trial plots within a catchment was used for that catchment. Further details
about these trials are contained in Peel et al. (1995) .
3
Effect of vegetative cover on USDA curve numbers
The procedure to determine the cover-CN2 function involved optimising the CN2 over
short periods of time (duration of two months) within the growing season, when cover
and growth did not change dramatically. This resulted in a number of paired cover
optimised CN2 values for each catchment. The observed runoff available for the
optimisation, for all plots, was the measured runoff for a whole catchment managed
very similar to the plots (catchment 5 - no grazing). This catchment encompassed two
of the six. trial plots. This was, however, an inferior procedure as the same runoff
values were used for all the catchments, defying the inherent differences in the physical
conditions of the catchments. As a result, the optimised values of CN2 obtained using
this procedure were deemed dubious and were not utilised for obtaining the cover
CN2 function. The derived values ofCN2 are, however, presented in Table 1.
Table 1. CN2 for DDRF using Runoff from Segments of a Season
Period Cover CN2 Sample Predicted RMSE Int** Slope** r no. * Observed
(%) nun nun nun/nun
Catchment 3 01110/92-30/11192 54.0 54.0 11(1) - - - - -01112/92-31101193 66.0 30.0 22(6) 0.83 1.25 -0.19 1.16 0.87 01102/93-31103/93 86.0 40.0 15(5) 1.09 0.71 0.30 0.59 0.35
Catchment 4 01110/92-30/11/92 63 .0 49.0 11(1) - - - - -01112/92-31101193 74.0 30.0 ·22(6) 0.74 1.08 -0.17 1.05 0.87 01102/93 -31103/93 92.0 40.0 15(5) 1.08 0.58 0.17 0.79 0.55
Catchment 5 01110/92-30/11192 50.5 52.0 11(1) - - - - -01112/92-31101193 66.0 29.0 22(6) 0.98 3.08 -0.22 1.38 0.85 01102/93-31103/93 87.0 39.0 15(5) 1.11 0.88 0.24 0.43 0.19
* * Figures in parentheses in the fourth column are the number of runoff events greater than 0.0 mm (ie. , runoffproducing rainfall events)
* Linear Regression: Predicted = Slope x Observed (or Measured) + Int Cover stands for mean cover during period of simulation
In Table 1, it should be noted that for the period 01110/92 - 30/11192, there was only
one runoff event greater than 0.0 mm. Hence, there were no statistical details for this
period.
4
Effect of vegetative cover on USDA curve numbers
3.1.2. Using Runoff Data over a Season
Motha et al. (1995) presented optimised CN2 for five different cover conditions at the
Douglas Daly Research Farm (DDRF), and derived parameters for runoff computation.
Data obtained in the 1994/95 growing season (November 1994 to May 1995), from
five runoff plots (SCI, SC2, SC3, SC4 and SC5), each 20 m x 5 m in area, were used.
Runoff plot SCI (within catchment 3) was bare and scalded. SC2 (also within
catchment 3), SC3 (within catchment 4) and SC4 (within catchment 5) with improved
pasture (Urochloa mosambicensis) were managed appropriately to maintain different
covers in each plot, which remained fairly constant through the duration of the
experimental program. SC5 (within catchment 6) was a native woodland plot
(Eucalyptus foelceanaIHeteropogon contortus dominated). Runoff, cover and soil
moisture were measured fairly regularly in all the plots. Further details about these
experiments are found in Dilshad et al. (1994) and Motha et al. (1995) .
For the optimisation exercise, whilst good agreement between predicted and observed
runoff and soil moisture were looked for, the predicted vegetative mass too was
scrutinised to be within a reasonable (based on knowledge and experience) range. The
time-averaged observed cover for the period of simulation was checked to be nearly
equal to the predicted time-averaged cover. The resulting cover-CN2 function is
shown in Figure 1.
100
iii : .. -.-.. -~- .--.. - .. - .~--.-.. - .. ~. ---.-.-.~-.. -.. - .. -.. ~ --.- .. - .. -~.---.-.. -.~ .. -.. - .-.. -i- -.-.. - .. -. !: i. i i ! i ! . !;
N 60 ·· __ ······4-·"--·"-1-"·"·"·_+--"·" ·";----"·" : '_· ·····-+ ········· ··-I----······_+······ .. ·--t··_-····-...
~ 40 __ ._._.1._ .. _._ .. _1_ .. _ .. _ .. -.1 __ ._ .. _ .. '-___ ._._+ .. _ .. _ .. _.J_._._ .. ~. ___ ._._.l .. _ .. _._.~ __ ._ .. _ .. _.
80
y = -O.49x + 84.52 ~ ; 1
20 r2 = 0.37 ---·-·-1-"-·-·-"1--.. -.. -·-i·---.. -·-+·-.. ---.. t--·-.. -.. -! i ! ! ! !
O+-~~-+---r--+---r; --~; ---r--~--r; --~
o 10 20 30 40 50 · 60 70 80 90 100
Cover (%)
Figure 1. Cover-CN2 Function for Improved and Native Pastures at DDRF
5
Effect of vegetative cover on USDA curve numbers
In Figure 1, the optimised CN2 values for improved as well as native pastures have
been combined; hence the scatter. Note that CN2 for SCI serves as the pivotal point at
near-zero cover, thereby dissociating itself from the influence of pasture type. The
derived function possessed a value of 85 at zero cover, and decreased 0.49 units for
every 1 % increase in cover. Although CN2bare was felt to be lower than was expected
(based on the study ofDilshad and Peel, 1994) cover had a pronounced effect on CN2.
Based on a general understanding of runoff on Australian soils, it was deduced that
CN2 stays constant beyond 80% cover.
Due to the difference in vegetation as well as difference in the cultural history of the
plot, the CN2 for SC5 should be treated separately. Figure 2 gives the cover-CN2
function for improved pasture plots, all three of which have a history of previous
cultivation. CN2 was 91 when bare and decreased 0.52 units for a cover increase of
1%. These values look reasonable based on the study ofDilshad and Peel (1994).
100
80 "- " -"+---" - " - " -~'---" -'-+" - " - " -'-1"-'-" - " -+--_ ... _ .. !-..... _ ... +_ ........ . ! ! i
i ! ! i 60 . : ········-+---···-·,r·····-···+-·········
N ",.,' I I
~ 40 __ ._._.J Y = -~.52x ; 91 .03!--·_·-t··---EJ9f---·_·-t·_··_·_·t-·_··_·_·
20 r2 = 0.59 ... _ .. _._.l .. _ .. _ .. _.~ __ ._ .. _ .. _l_ .. _._._ .. l .. _ .. _._ .. 1 .......... _ f . . ! i
O+-~---+---r--+---r--+---r--~--r-~
o 10 20 30 40 50 60 70 80 90 100
Cover ('Vo)
Figure 2. Cover-CN2 Function for Improved Pasture Catchments at DDRF
In Figure 2, there is an obvious outlier; the optimised CN2 of 42 for SC4 was derived
from limited opserved runoff data. Although it is not desirable to obtain the cover-CN2
function from three data points, it is interesting to check the effect of omitting the CN2
value for SC4.
6
Effect of vegetative cover on USDA curve numbers
Figure 3 shows the cover-CN2 function for improved pasture catchments, derived
from optimised CN2 values for data sets covering the entire season. From a value of
88 when bare, CN2 dropped by 0.3 units for every 1% gain in cover. The cover-CN2
function in Figure 3 shows a strong relationship between CN2 and cover. However, it
requires further strengthening, using more data at other covers.
100
80 .. - · .. ·- .... · .. ··- · .... l· .. i ·· .. -.;;::·· .. ··-f·· .. !· .... ~ .. ·· .. ·f·.,.·· .. ~· ·-·· .. ·;·~ .... ··=: .. ·· .. ·· .. [! .... -=· ...... ··L'·· .. ··=· .. ·~[· .... = ....... . . ~ i ! N 60 ··-·"··"·i· .. ·_· .. ·+· .. · .. ·· .. ·';' .. _· .. · .. · .. i· .. .. _· .. ·· .. ·t·· .. · .. · .. ·l .. -· .. .. .. · .. L .. ··-· .. ·· .. ·· ~· · .. ··-· .. ··t-· .. ·· .. ·· .. · ~ j! ! ~! j
40 -_ ... _ .. j : .-- .-.. -.+ .. -.. -.. -.~--.. -.. -.. -~-.. -.-.-.. ~ .. -.. -.-.+ _ ..... ... . i i i
Y = .. 0.3Ox + 87.97 I ' r2 = 0.996 .--.-.. -~ .. -.. -.. -.. -! - - .. - .. - .. -~---.-.. -.+ .-.. -.-. ..t.-........ .
! 1 ! 20
i
O+-~r--+---r--+---r-~---r--~--~~
o 10 20 30 40 50 60 70 80 90 100
Cover (%)
Figure 3. Cover-CN2 Function for Improved Pasture Catchments at DDRF, without the Optimised CN2 for SC4
In treating the CN2 for SC5 separately, the CN2 for SC 1 once again serves as the
'pivot point' . The resulting cover-CN2 function presented in Figure 4 is very useful for
hydrologic analyses of undisturbed native pasture catchments on Alfisols in Douglas
Daly. It should be noted that a severe lack of information necessitated the need to use
only two points. In this case, from a value of 88 when bare, CN2 decreases by 1.14
units for every 1% increase in cover. This represents a dramatic decrease in runoff
potential with increase in cover, in undisturbed native pasture lands in the Douglas
Daly region.
7
Effect of vegetative cover on USDA curve numbers
100
80 .. _._._ . _ .. _._ .. -.-.. _ .. _ .. _ ..... _._ .. _ ..... _-_._._.;_ .. _ .. _ .. _ ..... _._ .. _ .. -;._-_ .. _._ . .;. .. _ .. _._ . .l..-._ .. _ .. _. i ! !, 1. i . . ! . ! I I j iii ! i
60 -_···_··4····_···-t-···· .+ .. -.-.. -.. +.---.-.- .!-.. -.. -.. -.~--.-.. -.. -~.---.-.-·.f.··· ·· _···4--····· ·· · · ~ ! l ! !!!! o j i .;;; i
40 --···-·'i·········-t-.. _·······t--······'l'···-4i'-·j-···········i,,-········_}·_-_·······t·····_····t··_·_ .. ····· ! : I : ~ 1 j I !
20 --.-.-.. ~,' .-.. - Y =-1.14x +90.41 -L-.. ~-.. .l..-.-. ..l.---.-.-L-.-.-.L._ .. _._ i i '" i ! i I ! ~ :
O+-~---r--~--r--+; ---r--+-~-~!---r~ o 10 20 30 40 50 60 70 80 90 100
Cover ("10)
Figure 4. Cover-CN2 Function for Native Pasture at DDRF
3.2. Other Cover-CN2 Functions
Silburn and Freebairn (1992) using data collected during the period 1976-83 from
contour bay catchments in the eastern Darling Downs region of Queensland, for two
Vertosols (cultivated black earth soil and cultivated grey clay soil), studied the effects
of three fallow management strategies. They used the CREAMS hydrology model
(Knisel, 1980) and obtained optimised CN2 values for bare fallow, stubble mulch and
zero-tillage conditions, for the two cultivated cracking clays. It can be inferred from
their results that the value ofCN2 for zero cover is 73 and it drops 0.25 units for every
1% increase in cover. Littleboy et at. (1989) used the same data to build a sub-model,
that reduces CN2 with increase in cover, in their agricultural systems model PERFECT
(Silburn and Freebaim, 1992).
Littleboy et al. (1996) using data for the period 1989-92 on an Alfisol (cultivated hard
setting soil) in India, examined the effect of three different mulch treatments (bare,
farmyard manure at 15 tlha and rice straw at 5 tlha) and derived a cover-CN2 function.
The modified CNM with provision for the effect of cover on runoff was used within
PERFECT (Littleboy et at. 1989) to obtain the value of CN2 at different covers. CN2
for zero cover was found to be 95, which reduced by 0.35 units for every 1% increase
In cover.
8
Effect of vegetative cover on USDA curve numbers
Yee Yet (1994), using data collected during 1988-92 in Central Queensland, obtained
cover-CN2 function for Sodosols and Tenosols (both permanent vegetated hard
setting soils) under four different natural cover conditions. The optimised CN2 values
were obtained with the modified CNM coupled with the pasture growth model
GRASP (McKeon et aI. , 1993). CN2bare was determined to be 94; CN2 reduced by
0.525 units for each 1% increase in cover. Cover was deemed not to have any
influence on runoff beyond 80%.
The cover-CN2 functions reported in the above three studies are presented graphically
in Figure 5. An important factor that needs attention with optimised CN2 values
determined by Yee Yet (1994) is that, as opposed to daily comparison in all other
cases, the optimisation was based on observed runoff measured at the end of 'service
periods' . It was therefore, decided to check as an early exercise, the effect of using
runoff accumulated over a period of time, instead of daily runoff, on the cover-CN2
function.
- - - - - Cultivated hard-setting soil - - - - - - - Cultivated cracking clay
--Hard-setting pasture soil
100
90
80 70 .. .... . -.. .. .... .. .. .... .. .. .. .. .. .. .. 60 ---
N Z 50 (J
40
30 20
10 0
0 80 Cover (%)
Figure 5. Other Cover-CN2 Functions
3.3. Sensitivity of Frequency of Runoff Comparison on the Cover-CN2 Function
The data obtained from the runoff plots at DDRF in the 1994/95 growing season were
used to check the sensitivity of frequency of runoff comparison on the cover -CN2
9
Effect of vegetative cover on USDA curve numbers
function. Observed and predicted daily runoff accumulated over fortnightly intervals
was used to optimise CN2 for the five runoff plots. The optimised CN2 values were
exactly the same as before, although the optimisation statistics were inferior to those
for the daily comparison. There was no necessity to check the effect on the predicted
soil moisture as it is not an episodic attribute. The statistics for the two optimisation
exercises are summarised in Table 2.
SC
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
Table 2. Optimisation Statistics for Runoff Comparison on Daily and Fortnightly Basis
CN2 Sample Mean ProfJ.le Observed - Observed - Predicted - Predicted - Predicted no. Cover Depth Average StdDev Average StdDev Observed
mm mm mm mm mm
Soil Moisture
87 8 3 1500 434.00 27.00 467.00 46.00 1.08 73 8 52 1500 437.00 44.00 459.00 58.00 1.05 70 8 58 1500 411.00 47.00 431.00 52.00 1.05 42 3 63 1500 406.00 47.00 418.00 55.00 1.03 45 8 40 1700 514.00 53 .00 522.00 61.00 1.02
Daily Runoff
87 67 3 - 4.35 8.80 4.89 8.74 1.12 73 81 52 - 4.15 9.00 4.35 7.57 1.05 70 81 58 - 3.70 7.15 3.82 7.06 1.03 42 27 63 - 0.54 1.30 0.54 1.41 0.99 45 81 40 - 0.71 2.51 0.87 2.29 1.23
Fortnightly Runoff
87 8 3 - 36.46 16.27 40.91 14.49 1.22
73 9 52 - 37.37 28.96 39.17 39.34 1.05
70 9 58 - 33.32 16.44 34.38 39.79 1.03
42 3 63 - 4.90 1.15 4.83 4.14 0.99 45 9 40 - 6.37 6.31 7.83 12.67 1.23
4. Effect of Cover-CN2 Function on Hydrologic Predictions
r
0.60 0.80 0.94 1.00 0.85
0.90 0.68 0.57 0.73 0.05
0.45 0.97 0.70 0.98 0.04
Yee Yet (1994) and Motha et al. (1995) employed their respective derived cover-CN2
relationships in repeat simulations to check if there was any improvement in the
predictions. Both these analyses were performed within a pasture growth modelling
framework. The statistics for these repeat simulations are presented in Table 3.
10
RMSE
mm
43 .24 32.99 24.16 13.86 23.57
3.83 5.04 4.95 0.73 2.98
12.55 11.18 26.03
2.45 14.49
Effect of vegetative cover on USDA curve numbers
Comparisons reported by Yee Yet (1994), with the enhancements to runoff
computation, were good although there was no marked improvement from the
predictions using a constant CN2. Statistics presented by Motha et al. (1995),
however, indicated that the cover-CN2 function used in their repeat simulations
needed further scrutiny.
Table 3. Output Statistics of Repeat Simulations using Derived Cover-CN2 Functions
Pasture Type Site no. Sample Predicted RMSE Int** Slope** ~ no Observed
nun nun mmlnun
Central Queensland (Yee Yet, 1994)
Runoff deQth Native 1 15 1.05 8.6 0.0 1.0 0.7 Native 2 22 0.96 4.2 0.6 0.8 0.9 Native 3 19 0.85 16.9 -2.2 0.9 0.8 Native 4 30 0.98 13 .8 -7.7 1.2 0.9
Total soil moisture (profile depth = 500 mm) Native 1 10 0.99 6.57 16.0 0.83 0.99 Native 2 8 0.93 8.07 2.7 0.88 0.90 Native 3 9 0.97 6.38 4.4 0.98 0.90 Native 4 8 0.90 8.78 -2.3 0.94 0.83
Northern Territory (Motha et al. , 1995)
Runoff deQth Improved 5 67 0.99 3.84 0.09 0.99 0.77 Improved 6 81 0.61 6.25 0.75 1.35 0.58 Improved 7 81 0.56 5.50 1.43 1.09 0.46 Improved 8 81 2.81 4.54 0.22 0.26 0.20
Native 9 81 1.77 3.55 0.50 0.17 0.04
Total soil moisture (profile depth: Improved = 1500 mm, Native = 1700 mm) Improved 5 8 1.10 52.36 -178.1 1.51 0.65 Improved 6 8 1.07 38.85 2.34 1.06 0.83 Improved 7 8 1.07 30.79 -35.88 1.15 0.93 Improved 8 8 1.04 23 .91 -68.52 1.19 0.98
Native 9 8 1.01 23.49 -2.97 1.02 0.83
A close look at Table 3 reveals that the poor comparison of predicted and observed
runoff, obtained by Motha et al. (1995), is nearly proportional to the deviation of the
11
Effect of vegetative cover on USDA curve numbers
relevant optimised CN2 values from the cover-CN2 function (Figure 1). Hence, it
could be theorised that if the optimised CN2 values produced a close-fit for the cover
CN2 function, the predictions would be better.
5. Conclusions & Recommendations
The inclusion of the effect of cover on runoff, in the form of a cover-CN2 function,
makes the LAMS AT runoff sub-model physically more sound. In this study the
relationship between vegetative cover and CN2 for pastoral catchments in the Douglas
Daly region of the Northern Territory has been analysed. Similar studies reported by
earlier researchers have also been briefly discussed. The application of the derived
cover-CN2 function for DDRF in the simulation has not resulted in any noteworthy
improvement in the hydrologic predictions. This could be attributed to the difference in
the cultural history of the catchments.
It has been shown that delineating the optimised CN2 values based on the cultural
history of the catchments, yields superior correlation between the two runoff
parameters as opposed to the poor correlation obtained when all data were
amalgamated. In view of the extreme lack of knowledge in this avenue the delineated
cover-CN2 functions presented separately for previously cultivated pasture soils and
undisturbed native pasture soils are very valuable. The effect of the delineated cover
CN2 functions on the hydrologic outputs should be studied. However, information at
intermediate covers too are required to address the gaps present.
6. Acknowledgments
The help rendered by Luke Peel, Shane !zod, Jo Yee Yet, Mark Silburn, Greg
McKeon, Ken Day and staff at the Douglas Daly Research Farm is acknowledged.
Funds for this work, as part of the LAMSAT project, made available by the Land and
Water Resources Research and Development Corporation (LWRRDC), National
Landcare Program (NLP) and the Northern Territory Government, are also
appreciated.
12
Effect of vegetative cover on USDA curve numbers
7. References
Boughton, W.C. (1989). A Review of the USDA SCS Curve Number Method.
Aust. J. Soil Res. 27: 511-23 .
Dilshad, M., Motha, lA and Peel, L.l (1995). Preliminary Assessment of the
Influences of Pasture Cover on Surface Runoff, Bedload and Suspended Sediment
Losses in the Australian Semi-Arid Tropics. Conservation Commission of the
Northern Territory (CCNT) Technical Memorandum 94/12.
Dilshad, M. and Peel, L.J. (1994). Evaluation of the USDA Curve Number Method
for Agricultural Catchments in the Australian Semi-Arid Tropics. Aust. J. Soil
Res. 32: 673-85 .
Knisel, W. G. (1980). CREAMS, A Field Scale Model for Chemicals, Runoff and
Erosion from Agricultural Management Systems. USDA Conservation Research
Report No. 26.
Littleboy, M., Silburn, D .M ., Freebairn, D.M., Woodruff, D.R. and Hammer, G.L.
(1989). PERFECT: A Computer Simulation Model of Productivity, Erosion,
Runoff Functions to Evaluate Conservation Techniques. Queensland Department
of Primary Industries (QDPI) Brisbane Bulletin No. QB89005 .
Littleboy, M., Cogle, AL., Smith, G.D ., Yule, D .F. and Rao, K.P .C. (1996). Soil
Management and Production of Alfisols in the Semi-arid Tropics, India, I.
Modelling the effects of surface cover and tillage on runoff and erosion.
Aust. J. Soil Res. 34 (11): 91-102.
McKeon, G.M . Littleboy, M ., Carter, l and Flood, N . (1993). GRASP: Grass
Production Model. Draft Report, QDPI Brisbane.
Motha, lA, Dilshad, M. and Peel, L.J. (1995). Predicting Vegetative Cover, Runoff
and Soil Moisture for Assessing Land Degradation in Australia's Northern
Territory. Proceedings International Congress on Modelling and Simulation
MODSIM95 . 1:361-365.
Peel, L.l , Dilshad, M. and Motha, lA (1995). Results from SWIFTSYND Trials on
Improved and Native Pasture at Douglas Daly, N .T. CCNT Technical
Memorandum 95/02.
Silburn, D.M. and Freebairn, D .M. (1992). Evaluation of the CREAMS Model III.
Simulation of the Hydrology of Vertisols. Aust. J. Soil Res. 30: 547-64.
13
Effect of vegetative cover on USDA curve numbers
Williams, IR. and LaSeur, W.v. (1976). Water Yield Model using SCS Curve
Numbers. Am. Soc. Civ. Eng. J. Hyd Div. 102: 1241-53 .
Yee Yet, ID.S. (1994). Improved Runoff Prediction in Pasture Growth Models.
Project Report, Faculty of Engineering and Surveying, University of Southern
Queensland, Toowoomba, Qld.
14