Computers and Electronics in Agriculture 99 (2013) 258–264

Contents lists available at ScienceDirect

Computers and Electronics in Agriculture

journal homepage: www.elsevier .com/locate /compag

A web application for cotton irrigation management on the U.S. SouthernHigh Plains. Part II: Application design

0168-1699/$ - see front matter Published by Elsevier B.V.http://dx.doi.org/10.1016/j.compag.2013.08.019

⇑ Corresponding author. Address: USDA-ARS Plant Stress and Water ConservationLaboratory, 3810 4th Street, Lubbock, TX 79415, United States. Tel.: +1 806 7235237; fax: +1 806 723 5272.

E-mail address: Steven.Mauget@ars.usda.gov (S. Mauget).

Steven Mauget a,⇑, Gary Leiker a, Shyam Nair b

a U.S. Department of Agriculture-Agricultural Research Service, USDA Plant Stress and Water Conservation Laboratory, Lubbock, TX 79415, United Statesb Department of Agricultural and Applied Economics, Texas Tech University, Lubbock, TX 79409, United States

a r t i c l e i n f o

Keywords:CottonIrrigationDecision support toolsProfit estimationOgallala Aquifer

a b s t r a c t

A web-based application to help Southern High Plains cotton producers estimate profitability under cen-ter pivot irrigated production is described. The application’s crop modeling and general profit calculationapproach are outlined in a preceding companion paper, while additional details of the profit model andthe application’s operational features and GUI design are presented here. In addition, the assumptionsand approximations made in the application’s crop modeling and profit calculation are summarized,and directions are provided for accessing the application on the Ogallala Aquifer Project web page.

Published by Elsevier B.V.

1. Introduction 2. Functional overview

Dropping Ogallala Aquifer levels and volatile commodity pricesand energy costs make irrigation management an important butuncertain issue to cotton producers on the U.S. Southern High Plains(SHP). For example, is deficit or full irrigation more profitable underthe current pumping cost and lint price conditions? Also, what isthe most profitable way to divide production into dryland and irri-gated acreage with limited well capacity? To help SHP cotton produc-ers answer these questions the web application described hereestimates the effects of irrigation on the profitability of center pivotcotton production. This application’s main purpose is to estimatethe irrigation effect on yield and the related effects on both profitsper acre and profits over a center pivot area with combined drylandand irrigated production. A preceding companion paper (Maugetet al., 2013a, hereafter, Part I) describes the application’s foundations,i.e., the weather data used, the crop modeling procedures, yield cali-bration, and profit calculation at both the acre and center pivot areaspatial scale. The current paper describes the application’s opera-tional features and external design. Section 2 provides a brief over-view of the application’s main functional flow, while Section 3provides a description of the application’s working details, JavaScriptsoftware components, and external graphical user interface (GUI) fea-tures. Section 4 provides summary, caveats for the application’s use,and directions for accessing the application on the Ogallala AquiferProject’s (OAP) web page.

Fig. 1 is a functional flow chart of the application, which con-sists of three main parts. The first component consists of JavaScriptarrays that hold the simulated dryland and irrigated lint yieldsgenerated by the CSM-CROPGRO-Cotton (hereafter, CROPGRO-Cotton) model of the DSSAT crop modeling suite (Hoogenboomet al., 2010; Jones et al., 2003). The second component consists ofJavaScript and PHP code that calculates seed yields from the data-base’s lint yield per acre values, and the corresponding profit peracre from both lint and seed yields. Those profit values per acre(p) are calculated for both dryland production and irrigated pro-duction at 12 total irrigation levels. The dryland and irrigated pdistributions displayed by the application reflect the commodityprice, production cost, and pumping cost conditions defined bythe user using the application’s slider controls. The JavaScript codeof the final component estimates values of profit over a center pi-vot area (P) under mixed dryland and irrigated production, basedon calculated dryland and irrigated p values and a user-definedcenter pivot area (Acpv) and central well capacity (F).

3. Application features

3.1. ‘‘Read This First!’’ tab

The application’s graphical user interface (GUI), pictured inFigs. 2–5, consists of three main tabs. The ‘‘Read This First!’’ tab(Fig. 2) is a scrolling informational HTML page that provides anintroduction and covers the key points of Part I. This includesdescriptions of the weather and crop evapotranspiration (ETc.) dataused, the CROPGRO-Cotton modeling and yield adjustment process,and how the application calculates profit over a unit area and over

JavaScript Lint Yield Arrays

(Fig. 3d)

GUI Seed Yield Display

Per Acre Profit Calculation

Mauget et al., (2013a) Eqs. 2,3,4,5

Y’ =Y ± ΔYl l

0 lb acre ≤ ΔY ≤ 200 lb acre-1 -1

Y = α × Y’ s l

1.40 ≤ α ≤ 1.80s

s

Lint Yield Adjust (Fig. 2c)

Seed-Lint Ratio Adjust (Fig. 2e) User-Defined

Yield Value & Production Cost Conditions (Fig. 3b,g)

Center Pivot Profit ( ∏ ) Calculation

Mauget et al., (2013a) Eqs. 9,10,11

GUI Display

(Fig. 4a)

(TI)I

D

Y’l

YS

Yl

GUI Display(Fig. 5a)

∏

∏(TI,F,A )CPV

50 gpm ≤ F ≤ 700 gpm

Well Capacity (F) Adjust (Fig. 4c)

A = 126 or 503 AcresCPV

Define Center Pivot Area (A ) (Fig. 4b)

CPV

User-Defined Well Depth, Motor Type,Pump and Motor Efficiency (Fig. 3h)

Y’l

(Fig. 3a)

GUI Lint Yield Display

Fig. 1. Flow chart of web application’s main components. Gray outlined elements indicate user controls marked in Figs. 3–5. Gray shaded elements indicate Figs. 3–5’s yield,profit per acre (p) and center pivot profit (P) display features..

Fig. 2. Screen shot of the application’s ‘‘Read This First!’’ scrolling informational tab.

S. Mauget et al. / Computers and Electronics in Agriculture 99 (2013) 258–264 259

the area of a center pivot. The descriptions of weather and crop ETdata and crop modeling are more qualitative than that found in PartI, while the description of profit calculation – reproduced here inSection 3.3 – provides more detail. The ‘‘Read This First!’’ tab alsoincludes a basic ‘‘road-map’’ for the application and instructionsfor use, e.g., the various ways the application’s slider controls canbe operated.

1 http://www.jqplot.com.

3.2. ‘‘Yield Distribution by Irrigation Level’’ tab

Fig. 3 is a screenshot of the ‘‘Yield Distribution by Irrigation Le-vel’’ tab. The jqPlot1 graphics display on the tab’s upper graph(Fig. 3a) contains thirteen ‘‘Bar and Whisker’’ (B&W) diagrams for

the distributions of modeled CROPGRO-Cotton lint yields. These dia-grams mark the minimum, maximum, 25th, 50th, and 75th percen-tiles of yield for dryland production and 12 irrigation levels. In theirrigated simulations 11.0–22.0 in. of total irrigation was applied inincrements of 1.0 in. As described in Section 2 of Part I, each distri-bution was formed from the results of simulations conducted withweather data from four SHP sites during 1975–2004. The drylandand irrigated simulated yields for each irrigation level were aggre-gated across the four sites, resulting in 120 yields per distribution.As a result, the lower and upper whiskers mark the range of the low-est and highest 25% of yield values, i.e., the lowest and highest 30

Fig. 3. Screen shot of the application’s ‘‘Yield Distribution by Irrigation Level’’ tab. (a) Bar and Whisker (B&W) distributions showing maximum, minimum, and 25th, 50th,and 75th percentiles of lint yields for dryland production and twelve levels of total May–September irrigation. (b) Pop-up display showing a B&W diagram’s correspondingpercentile values when the mouse is hovered over the diagram. (c) Slider control for adjusting the yield database’s irrigated lint yields. (d) As in (a) for seed yield valuesderived from lint yield values. (e) Slider control for adjusting the ratio of seed-to-lint yield..

260 S. Mauget et al. / Computers and Electronics in Agriculture 99 (2013) 258–264

values. The central bar’s bottom and top mark the 25th–75th percen-tile of yields, i.e., the range of the central 60 values. In Fig. 3a the 120lint yields per acre generated in the dryland simulations is graphedon orange, and the diagrams for the 12 levels of total May 16–Sep-tember 17 irrigation applied in the simulations are graphed in green.In Fig. 3a and similar B&W displays in Figs. 4a and 5a the distributionwith the highest median value is graphed in yellow. Hovering themouse over these B&W diagrams produces a pop-up box listingthe diagram’s minimum, maximum, and percentile values as shownin Fig. 3b. Using the slider control to the right of Fig. 3a’s upper graph(Fig. 3c) the irrigated lint yield values used to calculate profits peracre can be uniformly adjusted upward or downward by as muchas 300 lbs per acre. The corresponding B&W diagrams for seed yieldsper acre are plotted on the tab’s lower graph (Fig. 3d).

As explained in Section 2.1 of Part I, CROPGRO-Cotton simula-tions generate cotton seed yields per acre (Ys). The database’sunadjusted lint yields (Yl) are calculated as seed yield divided by1.62, which is the ratio of the 2010 regional averages of seed to lintyields over the Economic Reporting Service’s Prairie Gateway re-gion (ERS, 2011). In the application the seed yields used to calcu-late profits are in turn calculated from the application’s currentlint yield values via a variable seed yield multiplier (as). Usingthe slider control to the right of the seed yield distribution graph(Fig. 3e), the ratio of seed to lint yield can be adjusted from the de-fault ratio (1.62) to values between 1.4 and 1.8. As in all of theapplication’s Dojo Toolkit2 JavaScript slider controls, Fig. 3’s LintYield Adjustment and Seed Yield Ratio sliders are labeled and accom-panied by a small dark-on-light blue3 display that indicates the sli-der’s current setting.

2 dojotoolkit.org.

3.3. ‘‘Calculate Profits’’ tab

Fig. 4 shows screenshots of selected features of the ‘‘CalculateProfits’’ tab. At the top of that tab are two sub-tabs that displayB&W diagrams of profits per acre and profit per center pivot area.Fig. 4a shows the ‘‘Profit per Acre’’ sub-tab exposed, which displaysthe corresponding distributions of profit per acre for the lint andseed yield distributions of Fig. 3a and d, derived from the applica-tion’s current yield income and production cost settings.

Yield income and production costs per acre are defined by theuser via slider controls located on seven sub-tabs located belowthe display sub-tabs (Fig. 4b). Fig. 4b screenshot shows the GUIwith the ‘‘Yield Income’’ sub-tab exposed. In the application totalyield income per acre (Iy) is calculated as the sum of lint and seedyield income.

Iy ¼ Ys � Ps þ Y l � Pl ð1Þ

where Ys is the cotton seed yield (ton per acre); Ps is cotton seedsales price ($ per ton); Yl is cotton lint yield (lb per acre); and Pl

is cotton lint sales price ($ per lb).To simulate the income effects of unpredictable variation in the

sale price of lint and seed the application provides the option ofrandomly varying the selected Pl and Ps values. The amount of thisvariation can be set to 0%, 5%, 10%, 15%, or 20% of the currently setsale price using the two dropdown lists on the right hand side ofthe Yield Income sub-tab (Fig. 4c and d). Thus, for example, if theLint Yield Price slider is set to $1.00 per lb and Fig. 4c’s dropdowncontrol is set to ±10%, the Pl values used to convert lint yield valuesto corresponding yield incomes are randomly varied between

3 For interpretation of color in Fig. 3, the reader is referred to the web version ofis article.

th

Fig. 4. Screen shots of the application’s ‘‘Calculate Profits’’ tab. (a) Profit display sub-tabs. The ‘‘Profit per Acre’’ sub-tab (shown) shows maximum, minimum, and 25th, 50th,and 75th percentiles of profits per acre for the corresponding lint and seed yield outcomes of Fig. 3a and d, based on the application’s yield value, production cost, andpumping cost settings. (b) Sub-tabs holding slider controls for defining yield income (shown), input costs, custom costs, labor and repair costs, miscellaneous costs, pumpingcosts, and harvest and gin costs. (c) Drop down list for setting randomized lint value variability. (d) Drop down list for setting randomized seed value variability. (e) Button toreset current dryland production costs to initial default costs. (f) Button to reset current irrigated production costs to initial default costs. (g) Screen shot of the input costssub-tab. (h) Screen shot of the pumping costs sub-tab.

S. Mauget et al. / Computers and Electronics in Agriculture 99 (2013) 258–264 261

Fig. 5. Screen shot of the application’s ‘‘Profit per Center Pivot’’ display sub-tab. (a) Maximum, minimum, and 25th, 50th, and 75th percentiles of profits estimated over acenter pivot’s area for the corresponding per acre profit outcomes of Fig. 4a, calculated with center pivot area and well capacity parameters defined by the user. (b) Drop downlist for setting center pivot area consistent with a quarter section (126 acres) or full section (503 acres). (c) Slider control for setting center pivot well capacity. (d) Center pivotarea displays showing the number of acres that can be irrigated at each of the twelve irrigation levels, assuming continuous irrigation during May 16–September 17. (e) Pop-up display showing the number of irrigated or dryland acres when the mouse is hovered over the green or orange areas of the pivot area displays in (d). (f) Displays indicatingthe well flow rate consistent with the irrigation levels of (a) and the irrigated areas of (d). (For interpretation of the references to color in this figure legend, the reader isreferred to the web version of this article.)

262 S. Mauget et al. / Computers and Electronics in Agriculture 99 (2013) 258–264

$0.90 and $1.10 per lb. To consistently reproduce the effects ofprice uncertainty in the simulated profit outcomes, this randomi-zation assumes that producers at the four modeling sites would re-ceive the same prices for their yields in any year. Thus one randomprice variable is calculated for each year of the 1975–2004 simula-tion period, and that price is then used to covert that year’s simu-lated yield from each of the four SHP locations to fourcorresponding income values.

Profit per acre (p) is calculated as the difference between yieldincome per acre and production costs per acre (Cp).

p ¼ Iy � Cp ð2Þ

Production costs per acre are calculated as the sum of per acrecosts set by the user via slider controls on the remaining 6 sub-tabs:

� input costs (Cin);� custom costs(Ccust);� labor and repair costs (Cl&r);� miscellaneous costs (Cmisc);� pumping costs (Cpump), and,� harvest and gin costs (Ch&g).

These costs are initially set to default values defined by theTexas Cooperative Extension’s estimated cotton costs for drylandand center pivot production during 2011 (AgriLife, 2011). Theapplication’s slider controls can be reset to these initial valuesusing the orange ‘‘Reset Dryland Production Cost’’ and green ‘‘ResetIrrigated Production Cost’’ buttons at the bottom of the calculateprofits tab (Fig. 4e and f).

3.3.1. Input cost tabInput costs are the sum of fertilizer, seed, and fuel related costs.

Cin ¼ ðCn � NtotalÞ þ ðCs � RsÞ þ Ctech þ Cfuel ð3Þ

where Cn is the nitrogen cost ($ per lb); Ntotal is the total nitrogenapplied in the dryland and irrigated simulations (lb per acre); Cs

is seed cost ($ per thousand seeds); Rs = 48.86 is Number of seedsper acre applied in the simulations (in thousands); Ctech is seed techfee ($ per acre), Cfuel is machinery fuel costs ($ per acre).

On the input cost tab CN, Cs, Ctech, and Cfuel can be independentlyadjusted for irrigated and dryland acres via separate slider controls(Fig. 4g). As the total applied nitrogen and seeding rates were fixedin the simulations, Ntotal and Rs are not user adjustable.

3.3.2. Custom cost tabCustom costs are the sum of per acre costs for the custom appli-

cation of fertilizer (Cfa), insecticide (Cia), herbicide (Cha), and har-vest aids (Chaa).

Ccust ¼ Cfa þ Cia þ Cha þ Chaa ð4Þ

Like the input costs, the four custom costs can be independentlyadjusted for both irrigated and dryland acres. As a result, the orga-nization of the Custom cost tab (not shown) is similar to Fig. 4g’sinput cost tab, as is the layout of the labor & repair and miscella-neous cost tabs.

3.3.3. Labor & repair costs tabLabor and repair costs include labor (Clab) and machinery repair

costs per acre (Crep), and an irrigation system repair cost (Cisr) thatis proportional to the total irrigation level, i.e.,

Cl&r ¼ Clab þ Crep þ ðCisr � TIÞ: ð5Þ

Here Cisr is the irrigation system repair rate ($ per acre-in. of irriga-tion); and TI is the irrigation rate used to produce the correspondingyield. That is, one of the 11.0, 12.0, 13.0,. . ., 22.0 in. total May 16–September 17 irrigation values assumed in the irrigated croppingsimulations described in Part I.

3.3.4. Miscellaneous costs tabMiscellaneous costs per acre for dryland and irrigated acres can

be set via slider controls for insurance costs (Cins), interest costs(Cint), fixed expenses (Cfix) and boll weevil assessment fees (Cbwa).

Cmisc ¼ Cins þ Cint þ Cfix þ Cbwa ð6Þ

3.3.5. Pumping costs tabTotal pumping cost per acre (Cpump) is calculated as the product

of a TI level in acre-in. and the cost of pumping one acre-in. of

S. Mauget et al. / Computers and Electronics in Agriculture 99 (2013) 258–264 263

water from a user-defined water table depth and forcing it througha center pivot irrigation system (Cacre-in).

Cpump ¼ TI � Cacre-in ð7Þ

The pumping cost per acre-in is calculated as,

Cacre-in ¼ ½PE � DW � ðDw þ Eþ 2:31 � ðPmin þ PfrÞ�=ðlm � lpÞ ð8Þ

where PE is the energy price ($ per kW h); DW is work required toraise 1 acre-in. of water 1 foot (0.08532 kW h per acre-in per ft); Dw

is well depth in feet; E is elevation head of a center pivot irrigationsystem (15.0 head-ft.); Pfr is pressure losses due to friction (lbs perin.2); Pmin is minimum pressure requirement for a center pivot irri-gation system (lbs per in.2); lm is pump motor efficiency; and lp ispump mechanical efficiency.

The energy cost, lm, lp, Pmin, and Dw variables can be adjustedvia the pumping costs tab’s vertical slider controls (Fig. 4h). Asdemonstrated in Section 3.2 of Part I, motor efficiency can haveclear effects on per acre profitability. To simulate these effects,the application provides three options for the pumping system’smotor type: electric, diesel or gasoline. When the diesel or gasolinemotor type is selected, fuel costs set via Fig. 4h’s leftmost slider areconverted to corresponding PE values using 37.95 or 36.64 kW hper gallon energy conversion factors. Pressure losses due to frictionare calculated via the Hazen–Williams equation (Menon, 2004).

Pfr ¼4:73 � L � F1:85

C1:85 � D4:87 ð9Þ

where L is total piping length of a quarter-section (1500 feet) orfull-section (2800 feet) center pivot irrigation system; F is center pi-vot well flow rate (gallons per minute); C is roughness coefficient ofsteel irrigation pipe (140.0); D is irrigation pipe internal diameter(8 in.).

3.3.6. Harvest & gin cost tabHarvest and gin costs per acre (Ch&g) are calculated as the sum

of stripping and module, ginning, and bag, tie and classing costs.

Ch&g ¼ ðCs&m � Y lÞ þ ðCgin � YscÞ þ ðCbtc � YbaleÞ ð10Þ

where Cs&m is the stripping and module cost ($ per lb of lint); Yl islint yield (lb per acre); Cgin is ginning costs ($ per cwt of seedcot-ton); Ysc is seedcotton yield (cwt per acre) = Yl/(rto * 100.0); Cbtc isbag, tie and classing costs ($ per 480 lb bale); and Ybale is bale yield(480 lb per bales acre) = Yl/480.0.

Seedcotton yield per acre is estimated from lint yields assuminga turnout ratio (rto) of 0.28. Like lint yield and seed yield salesprices, the same harvest and gin costs are used to calculate profitsover dryland and irrigated acres. As a result, the harvest & gin costtab’s layout (not shown) is similar to that of the yield income tab(Fig. 4b).

3.3.7. ‘‘Profit per Center Pivot’’ tabFig. 5 shows the ‘‘Profit per Center Pivot’’ display sub-tab of the

‘‘Calculate Profits’’ tab. This tab displays distributions of total profit(P) resulting from dryland, fully irrigated, or mixed productionover a center pivot’s area. As outlined in Section 4 of Part I,Fig. 5a’s B&W diagrams represent the distribution of P values de-rived from the application’s current irrigated profit per acre valuesat each irrigation level (pI(TI)) and the dryland profit per acre (pD)resulting from the same modeling site and simulation year. To esti-mate total profit over the pivot’s area the application multiplieseach pI value by the area that can be irrigated at each irrigation le-vel (ATI), while the remaining area under the pivot is assumed to bein dryland production (AD). As calculated via Eq. (8) of Part I, the ATI

values are a function of both the total irrigation level and the cen-ter pivot well flow rate (F). As a result,

PðF; TIÞ ¼ pIðTIÞ � ATlðTIÞ þ pD � AD ð11Þ

Dryland areas are calculated as the difference between a user-defined center pivot area (ACPV) and the ATI areas calculated foreach TI level and the user-defined well flow rate. As Fig. 4a’s dry-land and irrigated p distributions at each TI level are formed from120 profit values, the resulting P B&W diagrams displayed on thistab also reflect the distribution of 120 estimates of profit over acenter pivot’s area. As in Fig. 4a, a P distribution for strictly dry-land production over the center pivot area (AD = ACPV, ATI = 0) is in-cluded as the orange B&W diagram in Fig. 5a. The diagrams forfully irrigated or mixed production are graphed in white, whilethe distribution with highest median P value is graphed in yellow.As illustrated in Fig. 3a, hovering the mouse over these B&W dia-grams, and those of Fig. 4a, causes a pop-up to appear that displaysthe diagram’s minimum, maximum, 25th, 50th, and 75th percen-tiles. Pivot area can be defined via a dropdown list for either quar-ter section (126 acres) or full section (503 acres) center pivots(Fig. 5b). The pivot’s central well flow rate can be varied between50 and 700 gallons per minute using the slider control on the tab’sright hand side (Fig. 5c).

Once the user selects a center pivot area and a well flow rate,the application calculates the area under the pivot that can be irri-gated at each of the twelve irrigation levels. Those areas, and thecorresponding dryland areas, are plotted in green and orange inthe center pivot area tokens located under each of Fig. 5a’s B&Wdiagrams (Fig. 5d). Hovering the mouse over these green and or-ange areas triggers a pop-up display showing the correspondingarea in acres (Fig. 5e). As explained in Part I, Section 4, the applica-tion, when necessary, displays an adjusted value of the user-se-lected well flow rate that is consistent with the 12 irrigationlevels assumed in the cropping simulations and the repair andpumping cost calculations of Eqs. (5) and (7). For example, apply-ing a flow rate greater than 417 gpm to a 126 acre area over thesimulation’s 125 day irrigation period would cover that area withmore than the highest 22.0 in. irrigation level. When ACPV is setto 126 acres and an F rate above 417 gpm is selected, the applica-tion displays the flow rates consistent with each TI level in thewhite-on-dark blue displays below the center pivot tokens(Fig. 5f). Under those conditions, Fig. 5f text boxes display the Frates that provide each TI level to the entire pivot area over the125 day irrigation period. An example of this behavior can befound in Fig. 6a of Part I. When a selected flow rate provides a TIlevel to an area smaller than the selected pivot area, that level isapplied to the irrigated area calculated from Eq. (8) of Part I, andthe rest of the pivot is assumed to be in dryland production. Inthose cases, the flow rate display for that TI level equals that se-lected on the flow rate slider control, as in Fig. 5c and f.

4. Summary and caveats for use

A web application for managing center pivot cotton productionon the U.S. Southern High Plains (SHP) was described here. As de-tailed in a preceding companion paper and illustrated in this pa-per’s Fig. 1 flow diagram, the DSSAT CROPGRO-Cotton productionmodel was used to generate distributions of modeled yield out-comes per acre, based on 30 years of primary daily temperatureand precipitation data from four SHP weather stations and deriveddaily wind, dew point, and radiation values. Those modeled yieldswere generated under un-irrigated (dryland) and center pivot irri-gated conditions under 12 irrigation treatments that increasedfrom deficit to full irrigation levels. The modeled lint yields werethen adjusted to agree more closely with mean regional drylandyields and the water vs. yield response function of comparable irri-gated field study results. Those lint and seed yield distributions aredisplayed in the application on the ‘‘Yield Distribution by Irrigation

264 S. Mauget et al. / Computers and Electronics in Agriculture 99 (2013) 258–264

Level’’ tab (Fig. 3). Using the modeled dryland and irrigated yieldvalues, the application calculates corresponding profit values peracre under yield value, production cost, and pumping cost condi-tions set via slider controls on various sub-tabs on the ‘‘CalculateProfits’’ tab (Fig. 4b, g, and h). The resulting profit distributions ofprofit per acre for dryland conditions and irrigated production ateach of the twelve irrigation levels are then displayed on the ‘‘Prof-it per Acre’’ display sub-tab (Fig. 4a). Then, based on those drylandand irrigated profit values per acre and center pivot area and wellcapacity parameters set on the ‘‘Profit per Center Pivot’’ display tab(Fig. 5b and c), the application estimates and displays distributionsof total center pivot profit (Fig. 5a) under 12 partitioning optionsthat divide the pivot area into dryland and irrigated production(Fig. 5d).

In the process of deriving profit outcomes from modeled yields,the application makes a number of generalizations and assumptions:

� As described in Sections 2.1 and 2.2 of Part I, the irrigatedlint yields generated by CROPGRO-Cotton were adjustedto agree with a water response characteristic of local fieldexperiments. As those experiments were conducted withvarious cultivars and soil types, their water response func-tion, and that of the adjusted model response, should beconsidered generally representative of the impact of irriga-tion on lint yields in the SHP region. Stated otherwise, theapplication does not account for the yield or profit effectsof any individual cultivar or soil type.

� The yield database assumes no crop abandonment. Thusthe application’s profit calculations do not account for thepossibility of insect or weather-related losses experiencedin actual production conditions.

� The irrigated yields in the yield database assumes continu-ous irrigation over a fixed 125 day period (May 16–Septem-ber 17). As a result, although irrigation timing during acrop’s development stages might have important yield(Doorenbos and Kassam, 1979; Kirda, 2002; Moutonnet,2002) and profit effects, the application simulates theeffects of irrigation applied uniformly over the course ofthe growing season.

� The approach used to estimate total profits over a centerpivot’s area (Eq. (10a), Part I) assumes that the irrigatedand dryland profits per acre calculated by the applicationrepresent spatially averaged profitability over the pivot’sirrigated and dryland areas.

� The application’s profit analysis only accounts for annualoperating profit. No attempt is made to estimate the effectsof capital depreciation or the opportunity cost of center

pivot cotton production, e.g., the profit effects of growingother crops or using the land for non-agricultural purposes.

On Apple desktops and laptops the web application has beentested on the Safari, Firefox, and Google Chrome web browsers. OnPCs using the Windows operating system the application has beentested on the Internet Explorer, Firefox, and Google Chrome brows-ers. If using Internet Explorer, the user must upgrade to versions 9 or10. The application can be found on the Ogallala Aquifer Project’sweb site at www.ogallala.ars.usda.gov/CottonIrrigationTool.

Acknowledgements

All figures were produced using Generic Mapping Tools (Wesseland Smith, 1995). Thanks to Dennis Gitz of the USDA Plant Stressand Water Conservation Lab for helpful advice regarding cottonphysiology. Mention of trade names or commercial products in thispublication is solely for the purpose of providing specific informa-tion and does not imply recommendation or endorsement by theU.S. Department of Agriculture. The USDA is an equal opportunityprovider and employer.

References

AgriLife (Texas A&M University AgriLife Extention Service), 2011. Texas crop andlivestock budgets. <http://agecoext.tamu.edu/resources/crop-livestock-budgets/by-commodity/cotton.html> (accessed 05.10.11).

Doorenbos, J., Kassam, A.H., 1979. Yield response to water. FAO Irrig. And Drain.Paper No. 33, FAO, Rome, Italy. p. 193.

ERS (USDA Economic Reporting Service), 2011. Cotton production costs andreturns per planted acre, excluding Government payments. <http://www.ers.usda.gov/Data/CostsAndReturns/testpick.htm>.

Hoogenboom, G., Jones, J.W., Wilkens, P.W., Porter, C.H., Boote, K.J., Hunt, L.A.,Singh, U., Lizaso, J.L., White, J.W., Uryasev, O., Royse, F.S., Ogoshi, R., Gijsman,A.J., Tsuji, G.Y., 2010. Decision Support System for Agrotechnology Transfer(DSSAT) Version 4.5 [CD-ROM]. University of Hawaii, Honolulu, Hawaii.

Jones, J.W., Hoogenboom, G., Porter, C.H., Boote, K.J., Batchelor, W.D., Hunt, L.A.,Wilkens, P.W., Singh, U., Gijsman, A.J., Ritchie, J.T., 2003. DSSAT cropping systemmodel. Eur. J. Agron. 18, 235–265.

Kirda, C., 2002. Deficit irrigation scheduling based on plant growth stages showingwater stress tolerance. In: FAO (2002). Deficit Irrigation Practices. FAO Waterreport NO. 22, FAO, Rome, Italy, p. 3–10.

Mauget, S.A., Leiker, G., Nair, S., 2013. A web application for cotton irrigationmanagement on the U.S. Southern High Plains. Part I: Crop yield modelling andprofit analysis. Comput. Electron. Agric., (submitted for publication).

Menon, E.S., 2004. Liquid Pipeline Hydraulics. CRC Press, New York, p. 312.

Moutonnet, P., 2002. Yield response factors of field crops to deficit irrigation. In:Deficit Irrigation Practices, FAO Water report NO. 22. FAO, Rome, Italy, pp. 11–15.

Wessel, P., Smith, W.H.F., 1995. New version of the Generic Mapping ToolsReleased. EOS, Trans. Am. Geophys. Un. 76, 329.