Itc.nl

ITC Journal 1997-3/4
Biophysical sustainability of land use systems
esting prospects for early warning applications and cropyield forecasting.
The “sufficiency” of land unit properties can be gauged by monitoringselected system parameters and matching these with values calculatedfor a rigidly defined “production situation”. Land use system analysis must account for the dynamics of the system. The complexity of actu-al land use systems defies dynamic analysis. “Simplified” systems, in Land use systems are dynamic systems; the specifica- which limitations such as nutrient deficiencies, weeds infestation, tions of both the land unit and the land utilization type pests/diseases, harvest losses and “Acts of God” are assumed remediedthrough appropriate management activities, can be handled but the cal- change over time. Modern crop growth models account culated plant (organ) masses reflect potential rather than actual system for the dynamic nature of land use systems by applying performance. They may be valuable nonetheless: defining land units in a procedure of numerical integration over time: calcula- terms of their most “relevant” land qualities ensures that much of the tions of crop growth and development are done for a variation in the performance of actual land use systems is found backin the analyses of simplified systems that have a reference function.
succession of short time intervals in the growing cycle Analyzing the “yield gap” between calculated reference and observed of the crop. Dependent “state” variables signify the actual systems performance, with maximum use of geo-information to state of the system during a particular interval. (Interval facilitate regionalization from point analyses to an analysis of land, is a lengths are normally 1 day: a trade-off between the way to examine the sustainability of actual land use systems. The requirements of system dynamics and the availability of approach is now being examined in cooperative research projects inChina and Zimbabwe.
The sequence of intervals starts with the moment of crop emergence or planting. The calculations for thefirst interval in the growing cycle start with known val- ues of all state variables. Interval-specific values of In its simplest form, a land use system is composed of exogenous “forcing” variables (eg, weather data and one land utilization type practised on one land unit.
management specifications) are called, after which Such “single” land use systems are core elements in land processes that take place are calculated assuming steady suitability assessment; procedures have been worked out state conditions for the duration of the interval. All state to quantify crop performance in simple land use systems variable values are updated at the conclusion of each set with uniform fields that are planted to pure stands of of interval calculations; their new values reflect the state annual food or fibre crops. More complex forms of land of the system in the next interval when new interval-spe- use can be handled as aggregations of single land use cific forcing variable values are called and calculations systems: rotations are sequences of single land use sys- of changes in state variable values are repeated. The tems, and intercropping can in theory (sic!) be analyzed sequence of interval calculations ceases when (the inter- by examining concurrent single land use systems that val of) physiologic crop maturity is reached or growth share the same land unit, provided that one accounts for has become impossible because of lethal temperature or the mutual competition for light, water and nutrients of This “state variable approach” befits the dynamic In the present context, a land use system is considered nature of land use systems but is prone to the rapid “biophysically sustainable” if the compounded sufficien- propagation of errors imported with the (frequently cy of relevant land attributes does not deteriorate under called) primary data and incurred in the calculations of the applied land use, and that against a realistic time such essential processes as assimilation, maintenance horizon. Sustainability is an equilibrium problem. Crop respiration and growth respiration. The procedure fol- growth modelling and monitoring of relevant “land qual- lowed is “transportable” provided that only universally ity indicators” [2] are the means to judge: valid chemical, physical and biologic laws are used in (1) the sufficiency of the system’s supply side— the algorithm and empirical (observed cause-effect) rela- defined in terms of land management attributes—in the tionships are avoided. Ideally, the interpretation proce- face of the compounded land use requirements (ie, the dure would be entirely time- and site-independent, and local effects would be brought in exclusively by time- (2) the sustainability of the system over the years.
and site-specific input data such as weather data and Comparing calculated (reference) crop production potentials with observed crop performance offers inter- State variable values such as the dry leaf mass or total plant mass per ha are calculated for any interval (day) in the growing cycle of a crop and are strictly system-spe- Wageningen Agricultural University and Department of Land Resource cific. In theory, they could be used as gauges of the Biophysical sustainability ITC Journal 1997-3/4
compounded “sufficiency” of all land qualities that incoming solar energy (radiation), the temperature, and define the land unit in the system (ie, the “supply side”).
the crop’s photosynthetic properties. In glasshouses, One would simply compare them with values calculated even light and temperature can be optimized, and pro- for a hypothetic system with similar specifications but duction is limited only by the properties of the crop.
with correctable limitations removed. Calculating such This explains why in Dutch glasshouses tomato produc- sufficiency values for a sequence of years would reveal tion reaches an incredible 50 kg/m/year or 500 the long-term biophysical sustainability of a particular land use system. This sounds too good to be true. It is The production calculated for a fully optimized pro- duction situation is normally greater than the productionrealized in commercial farming, and much greater thanthe output of subsistence farming; it is not the actual production but the biophysical production potential. The Real-world farming involves land use systems of con- “yield gap” between calculated production potential and siderable complexity. The “demand side” is composed observed actual production results from the compounded of a wide variety of “land use requirements”, such as effects of all limitations that confront a real-world “adequate light and temperature conditions”, “adequate farmer but that were supposed “corrected” in the opti- water supply”, “adequate nutrient supply”, “adequate mized/simplified system. (“Production gap” would be a weeding”, “adequate control of pests and diseases”, etc.
better term in this context: “production” refers to the Many of these requirements are too complex to be mod- total production of dry plant matter; “yield” denotes only elled. Consider, for example, the land use requirement the harvested produce and is normally a fraction of pro- “adequate supply of plant nutrients”, or—simpler—”ade- duction.) If considered in relation to the biophysical quate supply of nitrogen to the crop”. An analytical potential, the yield gap reflects the seriousness of all model of nitrogen supply to a crop would need to limitations in a land use system. It is thus a “land qual- describe the decomposition of soil organic matter (an important supplier of nitrogen to plants). It would need The biophysical crop production potential differs to describe the atmospheric deposition of nitrogen (with between years, in response to differences in available 10 to 40 kg N/ha/year not to be ignored) and the binding solar radiation and temperature. The same variations in of atmospheric nitrogen by symbiotic and autotrophic sunshine and temperature conditions also affect the actu- binders. It would have to account for nitrogen applied al (farmer’s) production but a positive correlation with manure and commercial fertilizers. It would need between calculated reference production and observed to describe nitrate losses by leaching and losses of actual production cannot always be expected. On the ammoniacal nitrogen through volatilization. It would contrary, in a situation with much sunshine and little have to describe interactions among all these processes, rainfall, the calculated production potential would be eg, because it is likely that nitrogen losses increase with high (high rate of assimilation under surmised optimum fertilizer application and that binders of atmospheric water availability), whereas actual production might be nitrogen decrease their activity after fertilizer use when sharply depressed by severe drought. For this and other metabolizable nitrogen is amply “available”. And the reasons, the (reference) biophysical production potential model would have to quantify the recovery of “avail- is frequently replaced by the “water-limited production able” nitrogen by the root system as a function of a potential”, ie, the production potential of a system in score of environmental variables. All this adds up to an which nutrient supply, plant protection and harvesting elaborate model with massive data needs, which would methods are optimized, and production and yield are probably generate results that were exceedingly expen- entirely conditioned by the sunshine, temperature and sive and not very accurate. And one might wonder actual water conditions over the growing period.
whether there was a need for such models if one can Note that the water-limited production potential is less remedy any problem of nitrogen deficiency by applying than the biophysical production potential in systems with an adequate dose of manure or commercial fertilizer that less than optimum availability of water. Under such is readily available and affordable. Similar considera- conditions, a smaller gap exists between the water-limit- tions apply to the modelling of (the consequences of) ed production potential and actual production than weeding, or the lack thereof; or of plant protection between the biophysical potential and actual production.
(largely conditioned by the population dynamics of indi- The difference between the two gaps represents the part vidual pests, which are in turn influenced by, inter alia, of the yield gap that was caused by sub-optimal avail- the weather that is “expected”). One can of herbicide or pesticide might eliminate the problem altogether! If you Calculations of the water-limited production potential want it straight: “It is impossible to describe low-input must keep track of the actual quantity of soil moisture farming accurately with analytical models. We can han- stored in the rooted surface soil at any moment in the dle neither the complexity nor the data needs.” crop cycle, and match “water availability” with “demand The situation is less complicated if one considers for water by the crop”. That is an added complication; high-input farming, and quite comfortable if one is deal- the simpler model of the biophysical crop production ing with the limited complexity of strongly simplified potential assumed the land quality “water availability to “production situations”, ie, of hypothetic land use sys- the crop” to be non-constraining. The greater data needs tems in which limiting (land unit) attributes are suppos- and the more elaborate algorithm needed to calculate the edly “corrected” through plant protection measures, fer- water-limited production potential are associated with tilizer use, irrigation or drainage. If all correctable lim- greater cost and greater errors. Sad but still acceptable: itations are indeed eliminated, a system’s biophysical the water-limited production potential is (normally) a performance would only be limited by the amount of realistic indicator of the biophysical possibilities of Biophysical sustainability
ITC Journal 1997-3/4
advanced farmers who can afford to correct “low-cost” (3) 10 consecutive years (1981 to 1990) of daily limitations with herbicides, pesticides and commercial weather data recorded at Quzhou weather station in the fertilizers. The water-limited production potential tends North China Plain. These files are called to fluctuate strongly between years (often more strongly “Quzhou81.dat” through “Quzhou90.dat”.
than the biophysical potential) and correlates better with After invoking the program by activating directory actual production than the biophysical potential does.
C:\PS123 and typing PS123 <Enter>, you are guided As a reference value, “water-limited production poten- through a series of questions on the screen. Read the tial” is generally preferred to biophysical production disclaimer on the intro-screen and press any key.
potential despite the greater data needs and error margin.
The next screen asks questions about the site and year Monitoring the total above-ground dry plant mass (eg, of weather data that will be used in the calculations (eg, through remote sensing) and comparing it with the (ref- Quzhou <Enter> and 83 <Enter> will make the program erence above-ground) production potential calculated for load weather data recorded at Quzhou in 1983).
a corresponding production situation would produce a The following screens invite you to specify the file time- and site-specific land quality indicator with direct that holds “your” crop data and to make a choice from relevance for early warning studies. It is even conceiv- the crops/varieties on file. For this demo version, you able that reference production potentials could be esti- type “crop.dat” <Enter> and choose either (1) “generic mated some time before the end of the crop cycle by maize” or (2) “generic winter wheat”. (Winter wheat substituting long-term average weather data for as yet and maize are grown in rotation in the North China unknown real data. Combining land quality indicators and estimated reference production might produce a You will then be asked if you wish to calculate (1) the workable crop yield forecasting procedure. The practi- biophysical production potential, PS1, or (2) the water- cal implementation of this hypothesis is now being stud- limited production potential, PS2. Select option (1).
ied in SAIL projects in Zimbabwe and China.
All biophysical data needed are now known to the program but some vital management information is stillmissing, ie, the day at which the seedlings emerge and the quantity of seed sown per ha. It is suggested that Publishing conference proceedings on CD-ROM has you do as the Chinese do: choose emergence of maize to the advantage that models can be added to a hypothesiz- take place around day 160 in the year (ie, sometime in ing paper like the one presented here. The crop produc- June) and answer that 25 kg of maize seeds are applied tion model and the (sample) data files prepared for this per ha. If you choose to grow winter wheat, you set purpose are “packed” in file PS123.ZIP on the CD- emergence at day 290 and apply 150 kg of wheat seeds ROM. Activating hyperlink “Install PS123” under the yellow button “Related Papers” on the CD-ROM will After all calculations have been done, a summary unpack all files and install them in two new directories: table (see Table 1) appears, which lists from left to C:\PS123 and C:\FAOCLIM. You need 2 Mb of free space on your hard disk to accommodate all files.
(1) DAY (ie, the day in the year, or in the next year if It is warmly recommended that you read file “demo.123” in directory C:\PS123. This (ASCII) file describes data file structures and model operation; it can (3) ECe (the electric conductivity of a saturated soil be copied to your printer or viewed with any text editor.
Detailed information on the structure of the model and (4) LIVSLEAF (living leaf mass; in kg/ha) on the functional relations used are given by Driessen (7) SSO (storage organs, ie, ears for wheat or cobs for (1) generic soil data file “soil.dat” (2) generic crop data file “crop.dat” TABLE 1 Summary table of constraint-free maize grown at Quzhou 1983
Production situation 1: MAIZE (generic file) is grown at Quzhou83 from DAY 160 onwards
DAY LAI Ece LIVSLEAF SROOT SSTEM SSO TDM CFWATER
160 0.01 0.00 8 8 4 0 20 1.00170 0.09 0.00 46 44 22 0 112 1.00180 0.44 0.00 261 213 116 0 590 1.00190 1.99 0.00 1094 776 479 0 2350 1.00200 4.22 0.00 2350 1544 1691 0 5585 1.00210 5.68 0.00 3180 1972 3872 73 9096 1.00220 5.56 0.00 3181 2059 5732 1120 12102 1.00230 4.50 0.00 2615 1951 6902 3704 15228 1.00240 3.35 0.00 1973 1820 6245 7382 17636 1.00250 1.75 0.00 927 1695 5634 9589 18772 1.00261 0.02 0.00 0 1574 5067 10069 18454 1.00 Biophysical sustainability ITC Journal 1997-3/4
(9) CFWATER (ratio of the actual and maximum tran- irrigation schedule. Be realistic: use between 3 and 8 spiration rates). The value of CFWATER indicates the cm of water per application and observe realistic irriga- effect of drought stress on assimilation and is, by defin- tion intervals (basin or furrow irrigation is applied in the ition, always equal to 1.00 under PS1.
area and sprinkler or drip irrigation is not an option).
Table 1 summarizes the growth of a maize crop (as Remember that Table 2 suggests that the first application defined by “crop.dat”) that was sown around 1 June of irrigation water would be needed around the 25th day 1983; emergence was on day 160 (10 June). Note that the crop reached maturity on day 261 (and can be har- If you have chosen to examine a winter wheat crop, vested some 10 days later after drying to a moisture con- you will be surprised with a summary table that presents tent of 12 to 15 percent). Assuming that some 85 per- the most important system specifications with 30-day cent of SSO is grain and that the grain has 12 to 15 per- intervals, rather than 10-day intervals as in the case of cent moisture at the time of harvest, the yield component summer maize. The program tries to accommodate the would be around 10 tons/ha. The total production of dry summary table on one screen. You’ll notice that the plant matter amounts to 18.5 ton/ha, which accords with crop reaches maturity on day 158 in the next year some 20 tons at 10 percent moisture.
(1984). Note the state of dormancy during the cold win- Table 2 summarizes the performance of the same crop ter and the resumption of growth in the spring of 1984.
if grown under rain-fed conditions. Choose to examine The calculated biophysical yield potential amounts to another scenario (same site, same crop), select the option some 7.2 tons per ha; the total dry mass is 13 tons per “water-limited production potential (PS2)” and answer ha, equivalent to some 13 to 14 tons of “above-ground the questions on soil information (file “soil.dat”; option plant matter” at “harvest moisture content”.
4: “Loams”) and on the initial electric conductivity (ECe; The model suggests that growing this wheat crop rain- set at 4 dS/m which suggests a non-saline soil). Choose fed might not be a good idea. If the initial soil moisture an initial soil moisture potential of 1000 hPa (or cm) and potential, surface storage characteristics, water table a surface storage capacity for water (ASSC) of 1 cm.
depth and salinity values are chosen as in the above sce- There is no water on the land at the time of emergence nario for maize, the crop fails (a “false start”). Starting (SSC = 0). The initial groundwater depth is 250 cm and 10 days later results in a calculated rain-fed yield poten- the option “F” (= “fixed”) applies forced drainage that is tial of a mere 300 kg per ha. You could remedy the sit- installed in the area at a depth between 250 and 300 cm.
uation by irrigating the crop: a sample run with eight Note that development of the rain-fed crop is faster applications of irrigation water (3 cm water of 1 dS/m than under PS1 (maturity is reached on day 257 rather on days 75 and 100, 4 cm on days 120 and 140, 5 cm on than 261 as shown in Table 1); the LAI values are less day 160, 7 cm on days 180 and 200, and finally 5 cm on (less assimilating leaf mass) and the productions of day 215 in the crop cycle) resulted in an expected water- “storage organ” (SSO) and “total dry mass” (TDM) are limited yield potential of 6.9 tons per ha. Note that this reduced to 3706 and 7363 kg/ha, respectively. The col- irrigation input of (a total of) 38 cm of water was insuf- umn CFWATER indicates that transpiration already lags ficient to lower the salt content of the rooted soil com- sharply behind the theoretical maximum value only three partment. One might try other scenarios, eg, with better to four weeks after emergence (CFWATER = 0.37 on quality irrigation water, or with smaller but more fre- The column ECe in Table 2 suggests that the electric Examining alternative irrigation schedules is just one conductivity (saturated soil extract!) of the rooted soil application of quantified production situation analysis.
compartment increases from 4.00 dS/m at emergence to You can change the weather specifications, soil specifica- 4.22 dS/m on day 257 (ie, mid-September 1983). This tions, crop specifications and any or all of the manage- increase might be undone by the next winter rains.
ment attributes. You can also substitute long-term aver- You might wish to run another scenario: answer the aged weather data for as yet unavailable measured data, question “Is irrigation applied?” with “Y” and define an and calculate “expected” yield potentials. The utilities TABLE 2 Summary table of rain-fed maize grown at Quzhou 1983
Production situation 2: MAIZE (generic file) is grown at Quzhou83 from DAY 160 onwards;
the plot is on Loams (PSIinit = 1000 hPa) and is 0 times irrigated; initial Ece is 4 ms/cm
DAY LAI Ece LIVSLEAF SROOT SSTEM SSO TDM CFWATER
160 0.01 4.00 8 8 4 0 20 1.00170 0.09 3.97 46 44 22 0 112 1.00180 0.44 4.07 261 213 116 0 590 1.00190 1.99 4.05 1094 776 479 0 2350 0.37200 2.31 4.08 1242 947 698 0 2887 0.12210 1.93 4.12 1088 916 842 26 2871 0.95220 2.05 4.20 1179 966 2212 909 5282 1.00230 1.56 4.21 893 906 2743 2948 7585 0.55240 0.89 4.21 497 845 2480 3859 8020 0.53250 0.00 4.22 0 787 2239 3871 7695 0.00257 0.00 4.22 0 754 2105 3706 7363 0.00 Biophysical sustainability
ITC Journal 1997-3/4
“selexion.exe” and “amdascon.exe” in the directory C:\FAOCLIM can help here. Read “demo.123” in direc- La “suffisance” de propriétés d’unité des terres peut être mesurée en tory C:\PS123 for instructions and a word of caution. Do contrôlant des paramètres sélectionnés de systèmes et en comparant keep in mind that the soil and crop data provided are ceux-ci avec des valeurs calculées par une “situation de production” défi-nie de façon rigoureuse. Une analyse de système d’utilisation des terres default values. They can never replace calibrated/verified doit prendre en compte la dynamique du système. La complexité des values and are supplied only to illustrate the procedure.
systèmes actuels d’utilisation de terres défie l’analyse dynamique. Dessystèmes “simplifiés”, dans lesquels des limitations telles que des défi-ciences nutritionnelles, l’envahissement des mauvaises herbes,pestes/maladies, pertes de récoltes et “volontés de Dieu” sont assumés, remédiés au travers d’activités de gestion appropriée, peuvent être prisen main, mais les masses de plant calculé (organe) reflètent la perfor- The earlier observation that “achieving sustainability mance potentielle plutôt que celle du système actuel. Ils peuvent néan- is an equilibrium problem” is of general validity: the moins être valables: la définition d’unités de terres en termes de leurs precondition of “equilibrium” applies at all scales and to plus “pertinentes” qualités assure que beaucoup de variations dans la per-formance des systèmes actuels d’utilisation des terres se retrouve dans all aspects of land use. If sustainable land use is to be les analyses de systèmes simplifiés qui ont une fonction de référence.
achieved in practice, we cannot limit our attention to L’analyse du “déficit des récoltes” entre la référence calculée et la per-formance observée des systèmes actuels, en utilisant au maximum biophysical supply and demand. We cannot ignore, inter l’information géographique pour faciliter la régionalisation à partir d’analyse par points jusqu’à une analyse des terres, est une façon d’exa- We who study land use and land suitability for crop miner la durabilité des systèmes actuels d’utilisation des terres. Cetteapproche est actuellement examinée dans des projets de recherche en production focus our attention on the supply side. At the macro scale, prospects for supplying more foodusing less destructive methods are bleak. Unused land areas tend to have a marginal suitability for arable crop-ping (at best!), and increasing yields on existing farm La “suficiencia” de las propiedades de unidades de tierras se puede cali-brar mediante el monitoreo de una selección de parámetros de sistema y land requires resources that are not normally available la comparación de estos con los valores calculados a partir de una “situa- and knowledge that must be acquired in a process of ción de producción” rígidamente definida. El análisis de los sistemas de learning and experimentation. In the meantime demand uso de las tierras debe tomar en cuenta la dinámica de los sistemas. Lacomplejidad de los sistemas actuales de uso de las tierras desafía el aná- lisis dinámico. Sistemas “simplificados”, en los cuales limitaciones tales It is the author’s considered opinion that sustainabili- como deficiencia de nutrientes, infestación por malezas, pestes y enfer-medades, pérdidas de cosechas y “Actos de Dios” se asumen remediados ty cannot be hoped for if population growth is not mediante actividades apropiadas de manejo, pueden ser examinados, brought under control. Our present efforts in this direc- pero las masas de planta (órgano) calculadas reflejan el funcionamiento tion seem to be sadly inadequate. We are moving ever potencial del sistema más bien que el actual. Sin embargo, estos siste-mas simplificados pueden ser valiosos: la definición de las unidades de tierras en términos de sus calidades más “relevantes” asegura que unagran parte de la variación en el funcionamiento de los sistemas actualesde uso de las tierras se encuentra en el análisis de sistemas simplificados que tienen una función de referencia. Una manera de examinar la soste-nibilidad de sistemas actuales de uso de las tierras es mediante el análisisde la “brecha de rendimiento” entre la referencia calculada y el funciona- Driessen, P M and N T Konijn. 1992. Land-Use Systems Analysis.
miento actual observado de los sistemas, con un uso máximo de informa- Dept of Soil Science and Geology, Wageningen Agric Univ, 216 pp.
ción geográfica para facilitar la regionalización desde el análisis de pun- Pieri, C, J Dumanski, A Hamblin and A Young. 1995. Land Quality tos hasta el análisis de tierras. Este enfoque está siendo ensayado en Indicators. World Bank Discussion Papers No 315, Washington DC, proyectos cooperativos de investigación en China y Zimbabwe.

Source: http://www.itc.nl/~rossiter/Docs/PS123/ITCJ_1997_243_Driessen.pdf