U

N
C
O
R
R
E
C
T
E
D
P
R
O
O
F
1
3
4
5
6 Environmental Modelling & Software XX (2004) XXXXXX
7 www.elsevier.com/locate/envsoft
8
9 Note
10
ArcCN-Runo: an ArcGIS tool for generating curve number
11
and runo maps
12 Xiaoyong Zhan
a,
, Min-Lang Huang
b
13
a
Kansas Geological Survey, The University of Kansas, Lawrence, KS 66047, USA
14
b
Disaster Prevention and Research Center, National Cheng-Kung University, Taiwan, Peoples Republic of China
15 Received 16 February 2004; received in revised form 16 March 2004; accepted 29 March 2004
16
17 Abstract
18 The development and the application of ArcCN-Runo tool, an extension of ESRI@ ArcGIS software, are reported. This tool
19 can be applied to determine curve numbers and to calculate runo or inltration for a rainfall event in a watershed. Implemen-
20 tation of GIS techniques such as dissolving, intersecting, and a curve-number reference table improve eciency. Technical proces-
21 sing time may be reduced from days, if not weeks, to hours for producing spatially varied curve number and runo maps. An
22 application example for a watershed in Lyon County and Osage County, Kansas, USA, is presented.
23 # 2004 Elsevier Ltd. All rights reserved.
24 Keywords: Curve number; Runo; Watershed; Modeling; Hydrology; GIS
25
26
27 Software availability
29 Name of software 30 ArcCN-Runo
31 Developers 32 Xiaoyong Zhan and Min-Lang
33 Huang
34 Contact address 35 Kansas Geological Survey, The
36 University of Kansas, 1930
37 Constant Ave., Lawrence, KS
38 66047, USA
39 Tel.: 40 +1-785-864-2173
41 Fax 42 +1-785-864-5317
43 E-mail 44 xyz@ku.edu
45 Availability 46 http://arcscripts.esri.com/
47 details.asp?dbid=13311 48
49 1. Introduction
50 In hydrology a curve number (CN) is used to deter-
51 mine how much rainfall inltrates into soil or an aqui-
52 fer and how much rainfall becomes surface runo. A
53 high curve number means high runo and low inl-
54 tration (urban areas), whereas a low curve number
55 means low runo and high inltration (dry soil). The
56 curve number is a function of landuse and hydrologic
57 soil group. The Soil Conservation Service (SCS; now
58 Natural Resources Conservation Service (NRCS))
59 curve-number method is the most common method for
60 predicting storm runo volume. Many watershed mod-
61 els such as AGNPS (Young et al., 1987), EPIC (Wil-
62 liams, 1995), SWAT (Arnold et al., 1996), and WMS
63 (http://www.ems-i.com/WMS/WMS_Overview/wms_
64 overview.html) use this method to determine runo.
65 Sediment and nutrient transports are then calculated
66 based on the runo. Some applications of this method
67 with watershed models in Kansas were reported
68 recently (Bhuyan et al., 2002, 2003; Tsou and Zhan,
69 2004; Misgna et al., 2004).
70 Traditionally, an area weighted average curve num-
71 ber for the entire watershed is used to study the runo
72 of a watershed. The details of the spatial variation in
73 the watershed are often lost. An ArcGIS tool, named
74 ArcCN-Runo, is therefore developed to facilitate
75 watershed-modeling work. Unlike raster mode,
76 ArcCN-Runo is designed for any shape of polygon in
77 order to keep irregular boundaries unaltered. Appli-
78 cation of dissolving techniques reduces processing time
79 signicantly. The curve-number database implemented
80 provides a exible way to use a reference table for the

Corresponding author. Tel.: +1-785-864-2173; fax: +1-785-864-

5317.
E-mail address: xyz@ku.edu (X. Zhan).
1364-8152/$ - see front matter #2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.envsoft.2004.03.001
ARTICLE IN PRESS
ENSO: Environmental Modelling & Software - ELSEVIER 27-04-04 13:11:52 3B2 Ver: 7.51c/W Model: 5 ENSO1500
DTD v4.3.1 / ENSO1500
U
N
C
O
R
R
E
C
T
E
D
P
R
O
O
F
81 curve number based on soil and landuse information.
82 Users can also develop their own database as simply as
83 to editing an Excel le. The tool can be used to design
84 and manage hydraulic structures and projects, to esti-
85 mate future discharges, and to predict watershed
86 response associated with changes in topography, soil,
87 landuse, and landcover (e.g. urbanization).
88 This paper documents the development of the
89 ArcCN-Runo and its application to a watershed in
90 Lyon and Osage counties, Kansas, for the estimation of
91 runo depth and volume based on the spatially varying
92 information on soil and landuse. The following sections
93 are organized in a step-by-step procedure through a
94 eld example to explain the development and appli-
95 cation of ArcCN-Runo. Fig. 1 shows the ow chart
96 for this software. The software was developed with Vis-
97 ual Basic 6 on an ArcGIS 8.3 platform and its structure
98 includes one clipping tool, one intersecting tool, one
99 working panel, and one curve-number database.
100 2. Processing data
101 Soil and landuse data are needed as inputs. In the
102 USA, these data are usually available on the internet.
103 The following procedures, as an example, explain how
104 to prepare data before using the ArcCN-Runo tool.
105 Soil data of two counties, Lyon and Osage, Kansas,
106 were downloaded from http://www.ncgc.nrcs.usda.
107 gov/branch/ssb/products/ssurgo/. These coverages
108 were projected from CGS_North_American_1983 to
109 NAD83 using ArcInfo. After adding these coverages
110 into ArcMap, a table called comp in soil data, which
111 contains the attribute hydrogroup, is linked to the
112 coverage for each country individually. These two cov-
113 erages for two counties were then merged with Geo-
114 processing wizard.
115 The landuse data of these counties were downloaded
116 from http://mapster.kgs.ku.edu/dasc/catalog/cor-
117 edata.html. First, the land coverage was added into
118 ArcMap exported as a shapele with appropriate pro-
119 jection, and then added into ArcMap again. The Kan-
120 sas GAP Analysis Land Cover database includes 43
121 land-cover classes for the state of Kansas. The attri-
122 butes covername in the landuse data are used.
123 After loading the ArcCN-Runo tool in .dll le for-
124 mat into ArcMap, soil and land data were processed
125 through the following three steps: (1) To focus on an
126 interesting area or a watershed, the soil and land data
127 for the watershed were clipped using a polygon feature
128 from the overlay layer such as a watershed-boundary
129 layer. (2) To reduce processing time, the soil and land-
130 use layers were dissolved before intersection, based on
131 the attributes hydrogroup in soil and covername
132 in landuse. For this example, the number of polygons
133 for soil was reduced from 611 to 4 and for landuse
134 from 4330 to 17 after being dissolved. Figs. 2 and 3 are
135 the dissolved soil and landuse layers, respectively. (3)
136 Soil and landuse layers were intersected to generate
137 new and smaller polygons associated with soil hydro-
138 group and landuse covername. This step keeps all the
Fig. 1. Flow chart of ArcCN-Runo.
Fig. 2. Soil data.
ARTICLE IN PRESS
ENSO: Environmental Modelling & Software - ELSEVIER 27-04-04 13:11:55 3B2 Ver: 7.51c/W Model: 5 ENSO1500
2 X. Zhan, M.-L. Huang / Environmental Modelling & Software XX (2004) XXXXXX
U
N
C
O
R
R
E
C
T
E
D
P
R
O
O
F
139 details of the spatial variation of soil and landuse, and
140 therefore it is considered more accurate than using ras-
141 ter grid to calculate runo or any average or dominant
142 methods to determine curve number.
143 3. Determining curve numbers
144 The curve number for each polygon was determined
145 from the soil and landuse information. All input and
146 output are stored in the newly generated layer after
147 intersection and the operation was performed through
148 the following panel (Fig. 4). An important step is to
149 match the covername of landuse to the landuse of
150 index table in the curve-number database. The curve-
151 number database was built based on the similar curve-
152 number databases of AnnAGNPS, Basins, and other
153 watershed models. For real eld application, the curve-
154 number database should be developed based on the
155 local eld data if possible. Users can prepare this data-
156 base easily by editing an Excel le and then exporting
157 it in .dbf format. The map showing spatial variation of
158 the curve number (Fig. 5) is generated automatically
159 after clicking on the button called nish match on
160 the working panel, Fig. 5. Using the statistics functions
161 in ArcMap, users can report immediately some statisti-
162 cal results of the curve number for the watershed.
163 4. Calculating runo
164 The runo is calculated based on the SCS curve-
165 number-runo method. The SCS curve-number
166 method assumes that, for a rainfall storm event, the
167 ratio of actual retention of soil after runo begins to
168 the potential maximum retention of soil is equal to the
169 ratio of direct runo to rainfall. This simplied
170 assumption (Ponce and Hawkins, 1996) results in the
171 following runo equation where the curve number
Fig. 3. Landuse data.
Fig. 4. Working panel.
ARTICLE IN PRESS
ENSO: Environmental Modelling & Software - ELSEVIER 27-04-04 13:12:06 3B2 Ver: 7.51c/W Model: 5 ENSO1500
X. Zhan, M.-L. Huang / Environmental Modelling & Software XX (2004) XXXXXX 3
U
N
C
O
R
R
E
C
T
E
D
P
R
O
O
F
172 (0 CN 100) represents a convenient representation
173 of the potential maximum soil retention (S):
runoff
rainfall 0:2S
2
rainfall 0:8S
if rainfall > 0:2S
runoff 0 if rainfall < 0:2S
1
174 where S 1000=CN 10 in inches and S 25 400=
175 CN 254 in mm, SI units.
176 Data provided by the Kansas Agriculture Statistics
177 Service (Policy Research Institute of The University of
178 Kansas, 2003) show that the average annual precipi-
179 tation from 1961 to 1990 for Lyon County is 38.66 in
180 (980 mm). The precipitation used for this example is
181 3:222 in 38:66 in/12 months. It can be interpreted as
182 the averaged monthly value or a single rainfall event by
183 assuming there are about 12 relative large rainfalls
184 annually. In short, it is a hypothetical precipitation
185 event for this example. Apparently, daily and event-
186 based runo can be calculated in a similar way.
187 Fig. 6 shows the runo in depth, responding to the
188 specied precipitation event. The runo in volume is
189 area multiplied by runo depth. The total area of the
190 watershed is 10
8
m
2
and the total runo in volume is
191 3 708 338 m
3
for the whole watershed.
192 5. Final remarks
193 The newly developed ArcCN-Runo tool and its
194 application to a watershed are reported to introduce
195 this tool to the research community for protection of
196 water resources and water quality in watersheds.
197 Improvements may be made through implementing
198 precipitation time series and considering additional fac-
199 tors, such as dry and wet antecedent moisture con-
200 ditions, aecting the determinations of curve number
201 and runo. They will be the main eorts for the future
202 version of this software.
203 6. Acknowledgment
204 Appreciation is expressed to Dave Young for match-
205 ing the covername in landuse in the watershed and
206 those in the database. XYZ appreciates the computer
207 support of the Chinese Academy of Sciences (Inno-
208 vation Project No. KZCX3-SW-428). The review com-
209 ments from Harry He, Department of Natural
210 Resources, New South Wales, Australia; Ming-Shu
211 Tsou, Hydrologic Inc., USA; and Li Zheng, Chinese
212 Academy of Sciences, were helpful in improving the
213 clarity of this manuscript. The authors would like to
214 thank Ms. Marla Adkins-Heljeson for helping to edit
215 this paper.
216 References
217 Arnold, J.G., Williams, J.R., Srinivasan, R., King, K.W., 1996.
218 SWAT: Soil and Water Assessment Tool. USDA-ARS, Grass-
219 land, Soil and Water Research Laboratory, Temple, TX.
220 Bhuyan, S.J., Kalita, P.K., Janssen, K.A., Barnes, P.L., 2002. Soil
221 loss predictions with three erosion simulation models. Environ-
222 mental Modelling & Software 17 (2), 135144, (http://www.scien-
223 cedirect.com/science/article/B6VHC-44NM3X4-3/2/b3c5b71bd1-
224 e8163ce823fda8327db976).
225 Bhuyan, S.J., Koelliker, J.K., Marzen, L.J., Harrington, Jr., J.A.,
226 2003. An integrated approach for water quality assessment of a
227 Kansas watershed. Environmental Modelling & Software 18 (5),
228 473484, (http://www.sciencedirect.com/science/article/B6VHC-
229 48CNX6P-2/2/9feb312b1de00f2201ef52325e7ade0c).
230 Misgna, G., Tsou, M., Zhan, X., 2004. Integrated GIS-AnnAGNPS
231 modeling system for watershed assessment. Journal of Hydrologic
232 Engineering (in review).
Fig. 5. Curve number generated. Fig. 6. Runo calculated.
ARTICLE IN PRESS
ENSO: Environmental Modelling & Software - ELSEVIER 27-04-04 13:12:18 3B2 Ver: 7.51c/W Model: 5 ENSO1500
4 X. Zhan, M.-L. Huang / Environmental Modelling & Software XX (2004) XXXXXX
U
N
C
O
R
R
E
C
T
E
D
P
R
O
O
F
233 Ponce, V.M., Hawkins, R.H., 1996. Runo curve number: has it
234 reached maturity? Journal of Hydrologic Engineering 1 (1), 1119.
235 Policy Research Institute of The University of Kansas, 2003. The
236 Thirty-Seventh Edition of the Kansas Statistical Abstract 2002.
237 Available from http://www.ku.edu/pri/ksdata/ksah/.
238 Tsou, M., Zhan, X., 2004. Estimation of runo and sediment yield in
239 the Redrock Creek Watershed using AnnAGNPS and GIS. Jour-
240 nal of Environmental Sciences (in press).
241 Williams, J.R., 1995. The EPIC model. In: Singh, V.P. (Ed.), Com-
242 puter Models of Watershed Hydrology. Water Resources Publica-
243 tions, Highlands Ranch, CO, pp. 9091000.
244 Young, R.A., Onstad, C.A., Bosch, D.D., Anderson, W.P., 1987.
245 AGNPS, Agricultural Non-Point Source Pollution Model: A
246 Watershed Analysis Tool. USDA Conservation Report 35.
247 USDA-ARS, Washington, DC.
ARTICLE IN PRESS
ENSO: Environmental Modelling & Software - ELSEVIER 27-04-04 13:12:33 3B2 Ver: 7.51c/W Model: 5 ENSO1500
X. Zhan, M.-L. Huang / Environmental Modelling & Software XX (2004) XXXXXX 5