Precision Agriculture and Sustainability

Precision Agriculture and Sustainability R. BONGIOVANNI rbongiovanni@correo.inta.gov.ar National Institute for Agricultural Technology (INTA), Manfredi, Co´rdoba, Argentina J. LOWENBERG-DEBOER lowenbej@purdue.edu Department of Agricultural Economics, Purdue University, 1145 krannert Building, West Lafayette, IN 47907-1145 Abstract. Precision Agriculture (PA) can help in managing crop production inputs in an environmentally friendly way. By using site-specific knowledge, PA can target rates of fertilizer, seed and chemicals for soil and other conditions. PA substitutes information and knowledge for physical inputs. A literature review indicates PA can contribute in many ways to long-term sustainability of production agriculture, con- firming the intuitive idea that PA should reduce environmental loading by applying fertilizers and pesti- cides only where they are needed, and when they are needed. Precision agriculture benefits to the environment come from more targeted use of inputs that reduce losses from excess applications and from reduction of losses due to nutrient imbalances, weed escapes, insect damage, etc. Other benefits include a reduction in pesticide resistance development. One limitation of the papers reviewed is that only a few actually measured directly environmental indices, such as leaching with the use of soil sensors. Most of them estimated indirectly the environmental benefits by measuring the reduced chemical loading. Results from an on-farm trial in Argentina provide an example of how site-specific information and variable rate application could be used in maintaining profitability while reducing N applications. Results of the sensitivity analysis show that PA is a modestly more profitable alternative than whole field management, for a wide range of restrictions on N application levels. These restrictions might be government regulations or the landowner’s understanding of environmental stewardship. In the example, variable rate of N maintains farm profitability even when nitrogen is restricted to less than half of the recommended uniform rate. Keywords: sustainability, environment, GPS, VRT, Argentina Introduction The concepts of precision agriculture (PA) and sustainability are inextricably linked. From the first time a global positioning system was used on agricultural equipment the potential for environmental benefits has been discussed. Intuitively, applying fertilizers and pesticides only where and when they are needed, should reduce environmental loading. This paper will explore the realities of PA and sustainability. Exactly how can PA contribute to sustainability? Have the environmental benefits been measured? The paper will start with definitions of sustainable agriculture and precision farming. The next section will review research on the environmental impacts of PA. The last section will provide an example of how site-specific infor- mation and variable rate application could be used in maintaining profitability while reducing N applications. Precision Agriculture, 5, 359–387, 2004 � 2004 Kluwer Academic Publishers. Manufactured in The Netherlands. Sustainable agriculture The meaning of ‘‘sustainability’’ has been long debated. The term was originally used to refer to agricultural and industrial technologies that reduced or prevented the environmental degradation often associated with economic activity. Hartwick (1978) and Solow (1974) defined it economically as the ability to maintain constant con- sumption or productivity by substituting between natural resources and manmade capital in production. In this context ‘‘manmade capital’’ encompasses anything developed by human effort, including both physical capital (e.g. equipment, struc- tures) and intellectual capital (e.g. information, knowledge). Pearce and Atkinson (1993, 1995) defined it environmentally by stating that natural resources and man- made capital complement each other in a production process and as natural re- sources are the limiting factor of production, they must be preserved. In 1972, the United Nations defined sustainability in a more general sense as ‘‘...aimed to meet the needs of the present without compromising the ability of future generations to meet their own needs’’ (WCED, 1987). More recently, sustainability has been associated with a holistic consideration of the economic, environmental, and sociological impacts of any development (Caffey et al., 2001) (Figure 1). Applying the concept to agriculture, the American Society of Agronomy (1989) defines ‘‘Sustainable Agriculture as the one that, over the long term, enhances environmental quality and the resource base in which agriculture depends; provides for basic human food and fiber needs; is economically viable; and enhances the quality of life for farmers and the society as a whole.’’ Precision agriculture Site-specific management (SSM) is the idea of doing the right thing, at the right place, at the right time. This idea is as old as agriculture, but during the mechani- Figure 1. Sustainability as described by the intersection of three disciplines: ecology, economics and sociology. BONGIOVANNI AND LOWENBERG-DEBOER360 zation of agriculture in the 20th century there was strong economic pressure to treat large fields with uniform agronomic practices. Precision farming provides a way to automate SSM using information technology, thereby making SSM practical in commercial agriculture. PA includes all those agricultural production practices that use information technology either to tailor input use to achieve desired outcomes, or to monitor those outcomes (e.g. variable rate application (VRA), yield monitors, remote sensing). Lowenberg-DeBoer and Swinton (1997) define SSM as the ‘‘electronic monitoring and control applied to data collection, information processing and decision support for the temporal and spatial allocation of inputs for crop production.’’ They high- light that the focus is on agronomic crops, but the arguments apply to horticultural crops and to the electronic tagging of livestock. Temporal SSM requires management of inputs based on information about the life cycles of agricultural crops, livestock or pests. This temporal information is often referred to as developmental stage (DS) information (Swinton, 1997). For instance, integrated pest management involves many cases of DS management practices, such as the use of pest scouting to determine the need and timing of pest control. DS management is also used in livestock management: bar-coding and other sensors are used to keep track of individual dairy cow milk production, food consumption, and health (Swinton, 1997). Ethical debate Agriculture cannot be sustainable if farmers use practices that are socially unaccept- able or not profitable. There are also good practical reasons to be concerned with a deteriorating climate, global change, excessive erosion, water pollution, and increasing resistance of pests to biocides. Such utilitarian concerns are enough for many to embrace sustainability as a goal (van Schilfgaarde, 1999). They are, in fact, the primary driving force behind the research done by the Water Quality and Management ARS National Program that targets PA’s effectiveness (Barry-Stelljes, 2000). Besides the utilitarian, physical aspects, however, there are philosophical and religious issues that deserve attention. One of these is stewardship. Sometimes stewardship of land is seen as a responsibility to future generations. In a religious context it is often seen as the responsibility to preserve and enhance God’s creation. In either case, land and nature in general is thought of as something that human beings are given a temporary responsibility to care for. This is in contrast to the view that sees natural resources as assets to be exploited for the personal gain of the current property owner. A related line of thought sees the farm as a living entity, an organism, and charges the farmer with the task of guiding this entity to produce crops and livestock in harmony with the environment. In some cases this view focuses on the farm as a self-contained entity with minimal dependence on purchased inputs and commodity markets. The objective of the farmer in this view is to enhance a biological balance, where the vitality of the land permits the harvesting of crops (van Schilfgaarde, 1999). PRECISION AGRICULTURE AND SUSTAINABILITY 361 One problem with the ‘‘farm as a self-contained organism’’ concept is the linkage between the farm, the community and the rest of the world. Unless farmers are to be hermits, they are part of a larger community and as part of that larger community they provide agricultural products and in return receive consumer goods and farm inputs made by their non-farm neighbors. The specialization of tasks allows all to achieve a higher standard of living. Whether those neighbors are mainly those in same village as was true in Medieval Europe, or scattered over the world in the globalized economy, nutrients are being exported from the farm and something must be imported to maintain the balance. Precision agriculture potentially provides producers improved tools to manage those inputs that must be brought to the farm. Instead of indiscriminately applying fertilizer or pesticides at uniform rates over large areas, PA allows producers to better target applications. It is often argued that PA substitutes information and knowledge for some external physical inputs, thereby potentially moving the farm closer to the ideal of biological balance. Of course, information technology and the knowledge that makes PA work are also external inputs. The hope of PA is that its use will be less disruptive of natural systems than uniform application of physical inputs has been. Challenges According to Hatfield (2000), a farming system is comprised of many elements, but the variations that exist within a field can be summarized in three classes of variation: (1) natural, such as soil and topography; (2) random, such as rainfall; and (3) managed, the fertilizer or seed application. The interaction among these three sources of variation results in offsite impacts. The natural variation includes: (a) soil variation, (b) biological variations, and (c) soil process variation (Hatfield, 2000). Soil varies spatially in water-holding capacity, organic matter, and other physical and chemical characteristics by topography, as well as by a series of interacting elements. The challenge is to quantify soil variation. Biological variations within fields are as great as soil variations, including soil microbial populations, weed populations, insect popula- tions, disease occurrence, crop growth, and harvestable yield, which is the variable that allows farmers to realize the outcome of all biological variations. Soil process variations are best understood by looking at N dynamics. One challenge is to quantify the response by varying response levels across soil types and topography, as Bongiovanni and Lowenberg-DeBoer (2001) did. Although the complexity of the interactions between the physical environment and the biological response creates a situation in which it is difficult to quantify the response to different practices, spatial regression analysis of yield monitor data, as related to soil characteristics shows promising results. Kachanoski and Fairchild (1996) illustrated the spatial scaling problem and the value of taking into account the spatial variability of fields. Their results suggested that since the relationships among yield response, soil test, and applied fertilizer are BONGIOVANNI AND LOWENBERG-DEBOER362 non-linear, a single soil test calibration cannot exist for fields with different spatial variability. Another challenge is to show that PA can have a positive impact on the envi- ronment. Unfortunately, only few studies deal with this objective directly, most of them arrive to that conclusion as a by-product of other studies (Hatfield, 2000). Such studies can be categorized as (1) nutrient management, (2) pest management, and (3) soil and water quality, and are summarized in Tables 2 through 6. Literature review of studies on nutrient management Schepers (1999), summarizes in Table 1 the environmental risks from nutrients and soil organic matter that are perceived to be the greatest for the different processes. The interactions between factors cited in Table 1 and processes must be addressed in any discussion of environmental quality. NO3–N losses are influenced by any factor that affects the movement of water within and from the field. This movement of N with water is believed to be one of the causes of hypoxia near the mouth of the Mississippi River, in the Gulf of Mexico, a condition in which water is depleted of its oxygen content, resulting in a serious reduction of biological activity (Hatfield, 2000). Hatfield (2000) also highlighted that the processes outlined by Schepers (1999) cannot be changed, but it is possible to modify the loading of nutrients and pesticides in a field, providing an opportunity for effective management of inputs through PA, while increasing production efficiency. Nitrogen (N) According to Wang et al., (2003), studies of economic and environmental impacts of variable N application in crop production have been mixed. They refer to studies that found Variable rate technology (VRT-N) to be superior to uniform rate in terms of economic and water quality benefits (Babcock and Pautsch, 1998; English et al., 1999; Schnitkey et al., 1996; Thrikawala et al., 1999). They also mention that in other papers, the benefits of VRT were not evident (Qiu and Prato, 1999; Watkins Table 1. Environmental risks from nutrients and soil organic matter Process N P K S OM Leaching + ) ) ) ) Denitrification + ) ) ) ) Eutrophication + + ) ) ) Precipitation + + + ) ) Runoff + + ) ) + Volatilization + ) ) ) ) Saltation ) ) + ) ) Source: Schepers (1999), as cited by Hatfield (2000). PRECISION AGRICULTURE AND SUSTAINABILITY 363 T a b le 2 . S tu d ie s o n th e im p a ct o f si te -s p ec ifi c N m a n a g em en t o n th e en v ir o n m en t C ro p In p u t/ F a ct o r R eg io n M et h o d o lo g y R es u lt s o f u si n g V R T W a n g et a l. ( 2 0 0 3 ) C o rn N M is so u ri U se d to p so il d ep th d a ta to d ev el o p re co m m en d a ti o n s w it h a si m u la ti o n m o d el . * V R T w a s m o re p ro fi ta b le th a n u n i- fo rm ra te in 7 5 % o f th e ca se s, w it h a g a in in p ro fi ts u p to $ 3 7 .1 4 h a ) 1 in o n e o f th e fi el d s. R o b er ts et a l. ( 2 0 0 1 ) C o rn N T en n es se e E P IC si m u la ti o n m o d el to es ti m a te N le a ch in g . * M o re N w a s a p p li ed w it h V R T , b u t le ss N w a s lo st to th e en v ir o n m en t N le ac h in g re d u ce d b y 2. 24 –4 .4 8 k g h a� 1 . D el g a d o et a l. ( 2 0 0 1 ) B a rl ey P o ta to N S o u th C en tr a l C o lo ra d o F ie ld tr ia ls a n d u se o f in fo rm a ti o n to es ti m a te N le a ch in g a s a d iff er en ce . * N m a n a g em en t p ra ct ic es ca n p o te n - ti a ll y b e im p ro v ed to re d u ce p o te n ti a l N lo ss es a n d co n se rv e w a te r q u a li ty . K h o ls a et a l. ( 2 0 0 1 ) C o rn N C o lo ra d o F ie ld tr ia ls a n d u se o f in fo rm a ti o n to es ti m a te N le a ch in g a s a d iff er en ce . * V R T -N h a s th e h ig h es t N U E a n d lo w es t le a ch in g co m p a re d to o th er tr ea tm en ts . W h it le y et a l. ( 2 0 0 0 ) P o ta to N W a sh in g to n S ta te F ie ld tr ia ls . M ea su re d N le a ch in g w it h p ro b es . * S u rf a ce so il h a d h ig h N O 3 – N fl u x . * S u b su rf a ce so il N O 3 – N fl u x st a b le . * N O 3 – N le a ch in g w a s d ec re a se d in v u ln er a b le zo n es d u e to a lo w er N ra te s in th es e zo n es . G ri ep en tr o g a n d K y h n ( 2 0 0 0 ) W h ea t B a rl ey N N o rt h er n G er m a n y F ie ld tr ia ls . M ea su re d re d u ce d ch em ic a l lo a d in g . * V R A re d u ce d N b y 3 6 % in lo w a re a s w h il e m a in ta in in g th e h ig h y ie ld s. E n g li sh et a l. ( 1 9 9 9 ) C o rn N W es t T en n es se e E P IC si m u la ti o n m o d el to es ti m a te N le a ch in g . * V R T w a s m o re p ro fi ta b le th a n u n i- fo rm ra te a n d th a t it g en er a te d le ss n it ro g en lo ss to th e en v ir o n m en t in m o st ca se s. BONGIOVANNI AND LOWENBERG-DEBOER364 R ej es u s a n d H o rn b a k er ( 1 9 9 9 ) N L a k e D ec a tu r Il li n o is E P IC si m u la ti o n m o d el to es ti m a te N le a ch in g . * V R T -N h a s th e p o te n ti a l to re d u ce th e m ea n a n d v a ri a b il it y o f N O 3 – N p o ll u - ti o n , w h il e im p ro v in g p ro fi ta b il it y . T h ri k a w a la et a l. ( 1 9 9 9 ) C o rn N O n ta ri o , C a n a d a S im u la ti o n m o d el (B a rr y et a l. , 1 9 9 3 ) to es ti m a te N le a ch in g . * N O 3 – N le a ch in g re d u ce d b y 1 3 % , a v er a g e o r b et w ee n 4 .2 % a n d 3 6 .3 % in h ig h a n d lo w fe rt il it y a re a s, re sp ec ti v el y . W a tk in s et a l. ( 1 9 9 8 ) P o ta to es N Id a h o E P IC si m u la ti o n m o d el a n d d y n a m ic p ro g ra m m in g to es ti m a te N le a ch in g . * N o en v ir o n m en ta l b en efi ts . * N o d iff er en ce s in N a p p li ed . * N o d iff er en ce s in N lo ss es . L a rs o n et a l. ( 1 9 9 7 ) C o n ti n u o u s C o rn N M in n es o ta L E A C H M si m u la ti o n m o d el to es ti m a te N le a ch in g . * N o 3 – N le a ch in g w a s d ec re a se d in 5 0 % , a v er a g e, o r fr o m 6 0 to 2 9 k g h a ) 1 . * D ec re a se w a s 0 k g h a ) 1 in th e lo a m , b u t 9 9 k g h a ) 1 in a lo a m y sa n d . L ei va et a l. ( 1 9 9 7 ) W h ea t R a p es ee d S o y b ea n s F er ti li ze rs P es ti ci d es S il so e E n g la n d M ea su re d ch em ic a l lo a d in g a n d es ti m a te d le a ch in g w it h si m u la ti o n . * P A le a d s to sa v in g s in fe rt il iz er s a n d p es ti ci d es , d ec re a si n g ri sk o f p o ll u - ti o n a n d en er g y u se , co n tr ib u ti n g to su st a in a b il it y . H er g er t et a l. ( 1 9 9 6 ) F u rr o w Ir ri g a te d C o rn N N eb ra sk a M ea su re d a ft er h a rv es t re si d u a l so il N O 3 – N a n d es ti m a te d N le a ch in g . * Im p ro v es N u se effi ci en cy . * R ed u ce s le a ch in g b y m in im iz in g h ig h N O 3 – N a re a s in th e fi el d . R ed u ll a et a l. ( 1 9 9 6 ) Ir ri g a te d C o rn N C en tr a l K a n sa s M ea su re d a ft er h a rv es t re si d u a l so il N O 3 – N a n d es ti m a te d N le a ch in g . * N o d iff er en ce s in N u se effi ci en cy . * N o d iff er en ce s in N O 3 – N le a ch in g . K it ch en et a l. ( 1 9 9 4 ) C o n ti n u o u s C o rn N M is so u ri M ea su re d g ra in p ro d u ct io n , u n re co v er ed N in th e cr o p , a n d p o st -h a rv es t N O 3 – N . * T h e a m o u n t o f u n re co v er ed N d ec re a se d in th e le a st p ro d u ct iv e so il s w it h V R T . * G ro ss sa v in g s o f $ 1 0 – $ 1 2 h a ) 1 fo r u si n g V R T . PRECISION AGRICULTURE AND SUSTAINABILITY 365 et al., 1998). In their research, Wang et al. (2003) evaluated the economic and water quality effects of adopting VRT-N and lime for corn production in Missouri. The methodology used topsoil depth data measured by soil electric conductivity, and developed fertilizer recommendations based upon a simulationmodel. VRT rates were compared to two different uniform N applications. Water quality benefits of VRT were evaluated based on potential leachable N. Results showed that VRT was more profitable than uniform rate in 75% of the cases, with a gain in profits up to $37.14 ha�1 in one of the fields. They also found that gre