effect of vegetative cover on usda curve numbers for ... · in the australian semi-arid tropics j....

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


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


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


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


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


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


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


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


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


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


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


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


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


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

effect of vegetative cover on usda curve numbers for ... · in the australian semi-arid tropics j....

Documents