TechnoGIN, a tool for exploring and evaluating resource use efficiency of cropping systems in east and southeast asia

In order to supportthe identification of sustainable land-use options and to support decision making with respect to land use, a tool was developed for quantifying inputs and outputs of c

TechnoGIN, a tool for exploring and

evaluating resource use efficiency of

cropping systems in East and Southeast Asia

Thomas C Ponsioen a,*, Huib Hengsdijk b, Joost Wolf c,*, Martin K van Ittersum d, Reimund P Ro¨tter c,

a

Agricultural Economics and Rural Policy, Wageningen University, P.O Box 8130,

6700 EW Wageningen, The Netherlands

bPlant Research International, Wageningen University and Research Centre,

P.O Box 16, 6700 AA Wageningen, The Netherlands

cAlterra, Wageningen University and Research Centre, P.O Box 47,

6700 AA Wageningen, The Netherlands

dPlant Production Systems, Wageningen University, P.O Box 430, 6700 AK Wageningen, The Netherlands

eNational Institute for Soils and Fertilisers, Chem, Tu Liem, Hanoi, Vietnam

fInternational Rice Research Institute, DAPO Box 7777, Metro Manila, Philippines

Received 18 February 2004; received in revised form 20 August 2004; accepted 29 November 2004

Abstract

Agricultural research in East and Southeast Asia is increasingly challenged by the searchfor land-use options that best match multiple development objectives of rural societies (e.g.,increased income, food security, and reduced environmental pollution) In order to supportthe identification of sustainable land-use options and to support decision making with respect

to land use, a tool was developed for quantifying inputs and outputs of cropping systems atthe field level TechnoGIN, the tool described in this paper, integrates systems analyticaland expert knowledge and different types of agronomic data enabling the assessment of inputs

0308-521X/$ - see front matter Ó 2005 Elsevier Ltd All rights reserved.

SYSTEMS

and outputs of a broad range of cropping systems and the evaluation of their resource use ciencies By using methods of spatial aggregation in combination with linear programming,results can also be used to explore trade-offs in resource-use efficiencies at higher levels such

effi-as the farm household, municipality and province New features in TechnoGIN comparedwith similar tools include the annual rotation of up to three crops, the distinction between aer-obic and anaerobic growing conditions of crops, and the procedure for estimating crop nutri-ent uptake TechnoGIN is illustrated with results from the Tam Duong district in NorthVietnam The design of TechnoGIN enables easy access to its data, parameters and assump-tions, and rapid generation and evaluation of input–output relationships of cropping systems

in order to add new information and to improve data TechnoGIN raises awareness about theassumptions incorporated and thus supports data collection and setting of the research agendawith respect to agro-ecological processes for which knowledge is incomplete, and is relevantfor showing trade-offs between production, economic and environmental impacts of differentland-use systems

Ó2005 Elsevier Ltd All rights reserved

Keywords: Land-use systems; QUEFTS; Resource-use efficiency; Rice-based systems; Systems analysis;

Linear programming

1 Introduction

East and Southeast Asia is increasingly challenged by various development tives of rural societies such as increased income, employment, improved natural re-source-use efficiency, food security, and reduced environmental pollution.Agricultural research therefore needs to be focused on the search for land-use op-tions that best match these objectives This calls for effective research tools enablingresource-use analysis at different levels of integration (i.e., farm household, munici-pality or district, province, and state) to support decision making with respect toagricultural land use These tools must be able to identify potential conflicts amongland-use objectives and resource use in order to generate technically feasible, envi-ronmentally sound, and economically viable land-use options that best meet awell-defined set of rural development goals

objec-Since the 1980s, the method of interactive multiple goal linear programming(IMGLP) has been proposed for an integrated analysis of resource use at regional

or farm level (De Wit et al., 1988) This method has been applied in various use studies (e.g., Van Latesteijn, 1995; Barbier, 1998; Bouman et al., 1999; Lu

land-et al., 2004) Key components in this approach are (1) databases on biophysicaland socio-economic resources and development targets, (2) a description of inputsand outputs of promising land-use activities, (3) a multiple criteria decision method(optimisation), and (4) sets of goal variables representing specific objectives andconstraints

This framework has been further improved and applied within the SysNet project,aimed at the development and evaluation of methodologies for exploring land-useoptions at regional scale in South and Southeast Asia (Hoanh and Roetter, 1998;

Roetter et al., 2005) Building upon this experience, a new research network, tems Research for Integrated Resource Management and Land Use Analysis in Eastand Southeast Asia (IRMLA)’’, has been set up for several multi-scale case studies inEast and Southeast Asia These studies combine the assessment of land-use alterna-tives with evaluation of stakeholder-negotiated choices at different decision levels(farm, district, and province) and supportive policy measures TechnoGIN, the tooldescribed in this paper, has been developed within this IRMLA project WithinIRMLA four case study areas have been selected: Batac and Dingras municipalities(Ilocos Norte province, Philippines), Pujiang county (Zhejiang province, China),Tam Duong district (Red River Delta, Vietnam), and O Mon district (MekongDelta, Vietnam).

‘‘Sys-TechnoGIN allows the quantification of inputs and outputs of large numbers ofcurrent and prospective cropping systems in these case study areas TechnoGINstands for Technical coefficient Generator for Ilocos Norte province, Philippines,

as it was originally developed for this province (Ponsioen et al., 2003) The term nical coefficient generator (TCG) is used for similar tools that were developed for thepurpose of explorative land-use analysis under multiple goals (De Koning et al.,1995; Hengsdijk et al., 1996, 1998, 1999; Bouman et al., 1998) The term Ôtechnicalcoefficient (TC)Õ refers to the inputs and outputs of land-use systems in economicand physical terms as quantified by this type of tool

tech-The purpose of this paper is to present the innovative aspects of TechnoGIN thatadd to the variety of approaches available TechnoGIN allows integration of differ-ent types of information on crop production and may support the scientific commu-nity in integrated analysis of cropping systems Important concepts that are used inTechnoGIN are defined in Section 2 The structure of the tool and its data require-ments are presented in Section 3 The calculation rules that were applied for nutrientand water balances, labour requirements and cost-benefit analyses, are presented inSection 4 To illustrate the type of output generated, an application is presented inSection 5 for the case study Tam Duong district In addition, application domains

of TechnoGIN output are indicated The new features of TechnoGIN comparedwith other TCGs, and factors that may affect the quality of its output, are discussed

in Section 6

2 Concepts

TechnoGIN enables the calculation of inputs and outputs of the so-called use systems (LUS), which are combinations of different land units (LU), land-usetypes (LUT) and production techniques Land units refer to areas of land that arerelatively homogenous in their biophysical (climate and soil characteristics) and so-cio-economic properties (input and output prices) Here, LUT is defined as a cropsequence of one, two or three crops per year Production techniques refer to the com-plete sets of inputs used to realise a well-defined yield (Van Ittersum and Rabbinge,

land-1997) In TechnoGIN, most inputs and outputs are calculated on a cropping seasonand an annual basis Exceptions are labour and water requirements, which are

expressed on a 10-day basis because the availability of both can be highly variable intime and may thus be decisive in trade-off analysis Besides marketable products andcrop residues, the undesirable outputs of cropping systems, such as soil nutrientdepletion and pollution of the environment by nitrate leaching and biocide emission,are calculated too.

TCGs are specifically developed to quantify differences in resource use of tional and improved land-use systems Hence, TechnoGIN enables us to analyse in-put–output relationships for both current and prospective cropping systems.Quantification of the relationships for current cropping systems is based on interpre-tation of survey data, whereas TechnoGIN simulates the information that is oftennot available from surveys, such as the amount of nutrients lost and water balances.Prospective or future-oriented cropping systems, however, are based on production-ecological knowledge, technical insight and required objectives, warranting increasedresource-use efficiency and yield levels as compared with those in current systems(Hengsdijk and Van Ittersum, 2002) Differences in efficiencies between productiontechniques can be ascribed to differences in farmersÕ management, knowledge (edu-cation), infrastructure (market for inputs and outputs), labour availability, etc.Key in calculating TCs for future-oriented cropping systems is the so-called Ôtar-get-orientedÕ approach implying that first a target output (i.e., yield) level is deter-mined, based on the biophysical conditions and the objectives for future cropproduction in the area under study Subsequently, the optimal combination of inputsrequired to realise this target yield is calculated with TechnoGIN This target-ori-ented approach enables us to quantify the minimum required amount of various in-puts such as labour, water, and fertiliser for a well-defined output In TechnoGIN,target yields are set equal to yields under Ôcurrent practiceÕ and Ôbest farmer practiceÕ,based on information from field surveys and experiments, literature, modelling, andexpert knowledge

conven-3 Model structure and input data

3.1 Structure and features

Similar to TCGs developed for West Africa (Hengsdijk et al., 1996) and CostaRica (Hengsdijk et al., 1998), TechnoGIN is programmed in Microsoft Excelwhereas all calculation rules are programmed in Microsoft Visual Basic for Applica-tions TechnoGIN consists of two files The main file contains the calculation rules, auser interface, and the generated TCs The database file must be created for eacharea under study and contains different types of data sets, organised into differentworksheets A simplified representation of the structure is shown inFig 1 This fig-ure shows the main parts of the system: (a) data bases, (b) user interface, (c) calcu-lations, (d) technical coefficients (i.e., the system output) The data bases in Excelsheets contain the required data described in Section 3.2 and listed in more detail

inTable 1 The user interface is described in the next paragraph The calculations

are described in Sections 4.1–4.5 The technical coefficients, or system output as ported to Excel or ASCI files, are also described in Sections 4.1–4.5 Examples ofoutput are given in Sections 5.4.1,5.4.2,5.4.3,5.4.4.

ex-Fig 1 Schematic representation of the structure of TechnoGIN The arrows represent flows of data Table 1

Data requirements per data sheet in TechnoGIN a

Data sheet Data requirements

Production techniques Relative nutrient use (R), biocide use (R) and water use efficiencies (R)

compared with those for current techniques Labour (R), fuel (R), machine (R) and animal use (R) proportionally to those under current techniques Prices of labour, fuel, machinery, draft animal and irrigation water (S) Crops Maximum yield (S or F), dry matter content (F), harvest index (F),

minimum and maximum N, P and K concentrations (F) in harvested products and crop residues, crop duration (S), crop coefficients (S), labour requirements per labour task (S), number of dekads needed for land preparation and harvesting (F), seed amount (F), fuel (S), machinery use (S), draft animal use (S), investments (S), recovery correction factor (F), anaerobic/aerobic (F), biocide use (S), farm gate prices (S), seed prices (S), current fertiliser rates for each land unit (S).

Land use types Crop rotation in one year (S), fraction of crop residues used as fodder,

burnt or mulched (S), low and high target yields per crop type and land unit (S).

Land units Long-term soil supply of N, P and K (S), maximum soil water holding

capacity (F), elevation and slope (S), fractions of sand, silt and clay (S), rainfall (S) and reference evapotranspiration (S) per dekad.

Biocides Active ingredient (S), duration (S), EPA/WHO index (S), and prices (S) for

each biocide type.

Fertilisers DM content (S), N, P and K concentrations (S) and prices (S) for each

fertiliser type.

Efficiencies Relative nutrient use (R), biocide use (R) and water use efficiencies (R)

proportionally to relative yield level.

Currencies Conversion rates (S) between different currencies for several years.

a

For each type of data, it is indicated whether its value is generally applicable and can be considered as fixed (F), whether its value should be established specifically (S) for each land use system, or whether its value is a relative fraction (R) which allows a rapid analysis of the effects (e.g., fertiliser demand) of relative changes in a factor compared with the standard value for a land use system (e.g., 20% more or less efficient nutrient use).

The friendly interface of the calculation file consists of buttons and forms required for database management and output analysis The buttons anduser-forms give access to data stored in the database file and allow rapid selection

user-of specific combinations user-of LUTs, LUs and production techniques After selection,TechnoGIN performs the input and output calculations of the required land-use sys-tems Generated TCs of cropping systems are stored in the main file as matrices andcan be exported to separate files, to be used in IMGLP models for further analysis.Various TCs can also be viewed in charts such as the monthly distribution of evapo-transpiration, crop water requirements and labour requirements per LUT Chartsare available showing different costs and economic returns of each generated crop-ping system facilitating cost-benefit analysis The calculated nutrient dynamics ofcropping systems are presented in a flow chart showing at a glance the nutrient flowsbetween different components for each crop in a LUT

3.2 Data requirements

Current data used in TechnoGIN are based on farm surveys, field experiments,literature studies, and expert knowledge TechnoGIN uses simple relationships tocalculate the use of biocides, labour, fuel, machines, draft animals and seeds fromthese input data Next, the corresponding economic costs are determined in cost-ben-efit calculations These input data sets require information from typical farmersreflecting the current practice in the defined cropping systems (current systems)and from outstanding farmers using improved techniques in the same study area

or in similar circumstances (future-oriented systems) More information about thesesystems that may differ in their productivity, resource-use efficiency and environmen-tal impact are given in Section 3.3.Table 1summarises the specific data requirementsfor TechnoGIN, organised into different worksheets (e.g., crop, land unit and fertil-isers) In this table, it is indicated which data can be considered universally applica-ble (e.g., nutrient concentrations per crop type) and which data should be specificallydetermined for each land-use system By using relative factors (Table 1), the techni-cal coefficients for a system can be easily varied for analysing the sensitivity of theland-use system and its output to changes in nutrient use efficiency, water use effi-ciency, and labour demand, for example Note that as the data requirements of Tech-noGIN are considerable, the system can also be applied if part of the data (e.g.,water and/or biocide use) are not (yet) available A Quickstart manual is availablefor more information on minimum data requirements for TechnoGIN and its initialapplication (see Availability of TechnoGIN and documentation)

3.3 Production techniques

TechnoGIN enables the definition of different production techniques such as rent systems and prospective systems with high target yields and possibly increasedresource-use efficiencies (future-oriented systems) Some inputs are substitutable,such as herbicides and manual labour for weed management, and the use of draftanimals and machines for field preparation Production techniques may differ in

cur-the use efficiency of resources For example, water-use efficiency depends on differentaspects of the applied irrigation technique, i.e., surface water or groundwater, sprin-kler or furrow irrigation, irrigation intervals and timing (Bouman and Tuong, 2001).Similarly, nutrient-use efficiency depends on the method of fertiliser application (e.g.,single or split applications and/or balanced nutrient applications (Witt and Dober-mann, 2002)).

As an example, qualitative characteristics of three production techniques are scribed inTable 2 Technique A represents the current mode of production Tech-nique B has the same yield level as technique A but the inputs (e.g., fertiliser use)are calculated in a target-oriented way based on yield level Production technique

de-C is also defined in a target-oriented way assuming a further improved system with

a higher target yield and an increased use efficiency of fertiliser nutrients and biocides.This requires improved farm management and mechanisation of farm operations

4 Calculations

The following calculation methods are described: nutrient balances (Section 4.1),crop nutrient uptake (Section 4.2), water balance (Section 4.3), labour requirements(Section 4.4), and cost-benefit analysis (Section 4.5) A complete overview of calcu-lation methods is given in the documentation of TechnoGIN (Ponsioen et al., 2003)

4.1 Nutrient balances

N, P and K balances are calculated in kilogram hectare
1for each crop in a LUT(i.e., annual crop rotation) The incoming and outgoing nutrient flows and those be-tween the different components of a LUS (inorganic nutrient pool, crop, animal andorganic nutrient pool) are illustrated in Fig 2 Crop nutrient uptake (U) resultspartly in removal of nutrients in harvested products (H) and partly in recycling ofnutrients in crop residues These recycled nutrients largely come through the inor-ganic pool available to the crop in the next season The efficiency of nutrient recy-cling depends on the type of applied crop residue management, which may be

Table 2

Characteristics of three different production techniques

Recovery fraction of applied fertiliser nutrients Calculated Standard Increased Labour requirements for crop management Current Increased Increased Labour requirements for other tasks Current Current Decreased

burning resulting in ash deposition (AD), ploughing in of the residues, and animaluse for fodder resulting in manure application (M) For each land-use system, thenatural nutrient inputs from soil mineralisation (S), wet deposition (WD) and bio-logical fixation (FL, SY) should be specified, which depend on location-specific con-ditions (soil, climate) and management.

Nutrient losses by leaching (L), denitrification (D), volatilisation (V) and fixation(X) are calculated as fractions of fertiliser application (F), manure application (M)and nutrient recycling These loss fractions are established on the basis of field con-ditions (e.g., soil texture, anaerobic or aerobic) and may be based on results fromrepresentative field trials One minus the loss fractions results in the recovery fraction(RF) of applied nutrients The fertiliser requirement of future-oriented cropping sys-tems (see end of Section 2) is calculated by balancing all flows in and out of the inor-ganic nutrient pool:

F¼U
S
SY
WD
FL

4.2 Crop nutrient uptake

Crop nutrient uptake is calculated using the QUEFTS approach (Janssen et al.,1990; Witt et al., 1999) for a specified target yield level The QUEFTS approach used

in TechnoGIN calculates N, P and K uptake assuming a balanced nutrient supplyfor the selected crop The calculated uptake of N, P and K is bound by two border-

lines describing the maximum dilution (D) and accumulation (A) of N, P and K in

the plant in relation to yield level (Fig 3: YND and YNA, etc.) At low yield levels,

calculated N uptake is near the YND line and at high yield levels (near Ymax) N take is approaching the YNA line The same applies for the other two nutrient ele-ments The two border lines are calculated from crop-specific minimum andmaximum N, P and K concentrations (Table 1)

up-Fig 2 Incoming and outgoing nutrient flows of a LUS and flows between different components of the LUS F, fertilisers; S, mineralisation from long-term soil supply; WD, wet deposition; FL, nitrogen fixation by free-living organisms; SY, symbiotic nitrogen fixation; L, N and K leaching; D, denitrification;

V, N volatilisation; X, Irreversible P and K fixation; B, burning; A, removal of animal product; H, harvesting; U, nutrient uptake by the crop; AD, ash deposition; M, mineralisation of crop residues and manure.

perco-Water balance calculations start after the dekad in which the smallest amount ofwater is to be expected in the soil Irrigation water requirements are calculated foreach dekad by subtracting ET and losses from water inflow (due to precipitation)and amount of available water in the topsoil at the beginning of the dekad A max-imum amount is specified for available water in the rooted topsoil layer (e.g.,AVAIL = 100 mm) Excess amounts of rainfall (after filling AVAIL up to maxi-mum) are lost by percolation to deep soil layers.

4.4 Labour requirements

Labour requirements are defined for four types of operations: (1) land tion, (2) crop establishment, (3) crop management, and (4) harvesting For each cropwithin a LUT, total crop duration and number of dekads needed for land prepara-tion and harvesting are specified The time needed for crop establishment is set at onedekad and the rest of the total crop duration is reserved for crop management La-bour requirements are calculated per dekad by dividing the amount of labour neededfor each of the four operations evenly over the dekads in which they take place

prepara-4.5 Cost-benefit calculations

Prices for different inputs such as labour, machinery and draft animal use, ent types of fertilisers and different types of biocides are specified in the input data

differ-Fig 3 Two borderlines indicating maximum dilution (D) and accumulation (A) of N (left), P (centre) and

K (right) in the plant in relation to yield level Lines apply to rice and are used in the QUEFTS approach for calculating N, P and K uptake for a specified target grain yield Maximum yield level for rice is

indicated by Ymax

files The costs of the specified LUS can be calculated from these prices and the culated input use The price for each crop product is also specified in the input file.Crop yields times corresponding prices give the economic benefits from the specifiedLUS These benefits minus total costs (including labour costs) give the net return andthe benefits minus the total non-labour costs give the gross return (Section 5.4.4).

cal-5 Application to the case study Tam Duong district

5.1 Case study area description

Tam Duong district (Vinh Phuc province, North Vietnam) is located upstream inthe Red River Basin (21°180–21°270N, 105°360–105°380E), about 60 km northwest ofHanoi The district covers almost 20,000 ha of which half is mountainous with alti-tudes between 100 and 1400 m above sea level and the other half flat to hilly Climate

is characterised by an annual rainfall between 1400 mm in the lower part and

2000 mm in the upper part of the district with more than 80% of the rainfall betweenMay and October (Fig 4) Temperatures range between 15 and 21 °C in January,and 26 and 33 °C in June to August

There are three seasons in the Tam Duong cropping systems: the dry season tween the end of January and May, the wet season between May and September, andthe autumn season between September and January Rice, peanut, tomato, cucum-ber and eggplant are the most common crops in the dry season; the most commonchoice in the wet season is rice A wide variety of vegetables, i.e., cabbage, tomato,cucumber, kohlrabi, chilli, soybean, peanut, maize, and sweet potato, are grown inthe autumn season

be-The region is characterised by a large surplus of agricultural labour With a ulation of 1,20,000, population density is very high (625 persons km
2), and thereare few off-farm employment opportunities Intensification of agricultural produc-tion has resulted in decreasing water quality Hence, policy priorities in the Tam

pop-Fig 4 Monthly mean rainfall (mm) and monthly mean minimum and maximum temperatures (°C) at the Vinh Yen station site (105°37 0 , 21°23 0 ) in Tam Duong (1992 and 2002).