You are here

Yield

Yield is a measure of crop production for a given land area, generally measured at the field scale. Yield typically focuses on a limited portion of the plant, such as the grain for row crops. However, in many cases farmers uses nearly all parts of the plant for various purposes. For this reason, we suggest taking into consideration all plant parts used by farmers, remembering that stover left if the field is often consumed by livestock. In many cases, it may be reasonable to focus on grain yield and stover.  The portion left in the field and not consumed by livestock plays an important role in nutrient cycling and this is measured in the next indicator, crop residues.

How to operationalize the metric

Method of data collection and data needed to compute the method:

Yield cuts are when a destructive harvest is carried out to measure production at crop reproductive maturity for a known area of land. For indeterminate reproductive plants this will require multiple harvests. For tree and bush species, yield can be estimated by using algorithms of stem (trunk) diameter at a specific height related to overall tree product per area under crown. The biomass produced can be partitioned by tissue type, such as grain and stover.

Yield cuts are common in agronomic trials and can be used in combination with household surveys. They are typically performed by randomly selecting a location from an experimental plot or from a plot on a farm and cutting the plants from a measured area (e.g., one square meter). In some cases, the entire harvest of a field with a known area will be measured. For guidelines on sampling a plot with quadrats and rating the vegetation within those plots, see Anderson and Ingram (1993,  section 3.1.2, which is copied in the Crop Residue Productivity Indicator under measurement method 1).

The biomass collected for grains are typically dried to standard moisture (e.g., 15.5% for maize in the U.S.) before being weighed. This can be accomplished by drying the grain in the sun or in an oven at low temperatures. Alternatively, the weight can be adjusted if the moisture content is known.

Unit of analysis:

The unit of analysis is the dry weight of plant biomass, grain, and stover per area of land. A measured value of 100 grams of grain per square meter is equivalent to 1 metric ton per hectare.

Limitations regarding estimating and interpreting:

Plant populations often vary across a single plot and this needs to be considered if sampling a portion of the plot for a yield cut. Plant population density is especially important when estimating yield by extrapolating the measured production of a few plants. Measuring plant population density is preferably done by counting the plants in a known area at harvest time. In some conditions, plant population density can be estimated by the seeding rate (seeds planted per hectare), but this assumes high germination rates and low plant mortality. Knowing the plant population density is also important for being able to interpret the reasons for yield changes or differences.

Method of data collection and data needed to compute the method:

Household surveys or field-based farmer surveys are opportunities where farmers are  asked to report their production for a given plot and the area where the crop was produced. During these agricultural surveys, total production from the farm for each crop grown may also be ascertained. The questions are mainly enumerated after harvest or at the end of the agricultural year. The section on the area planted is measured and for each field the area harvested is also collected per crop. 

Data on amount harvested are collected and the units (bags, ox-carts, baskets) are recorded by the enumerator. It is necessary to ensure that a conversion factor for the units is obtained (e.g., five 90-kilogram bags of maize). In this case, the weight can be converted to a common unit like kilograms. To prevent under-estimation of yield, enumerators should also ensure that for perennial crops or some annual crops, they ask about the amount of  unharvested planted crop still in the field.

Unit of analysis:

The unit of analysis is the farmer’s estimated weight per area, or volume per area.

Limitations regarding estimating and interpreting:

Farmer recall of the amount harvested may be reasonably reliable but are typically in local units (ox-carts, baskets, sacks), which then need to be accurately converted to kilograms.

In most contexts, farmer estimates of area are not very accurate (Fremont and Benson, 2011). Farmers may overstate their land holding because land is a source of prestige. In other cases, farmers may under report land owned for fear of being targeted for land redistribution. Measuring field size is important and can now easily be achieved by walking the boundary with a handheld GPS unit which can precisely calculate the area in hectares. 

Method of data collection and data needed to compute the method:

Crop models can be used to estimate yields for a wider range of environments, including edaphic (soil) or weather conditions not observed during the experiment.

Various crop models are commonly used for this purpose. Two of the most common are open-source software with strong support for global use, namely DSSAT (Decision Support System for Agrotechnology Transfer, http://dssat.net/downloads/dssat-v46) and APSIM (Agricultural Production Systems sIMulator, https://www.apsim.info/AboutUs.aspx). Peer reviewed publications are available describing these models (for DSSAT, see Jones et al., 2003; for APSIM see Holzworth et al., 2014). 

Crop models simulate crop growth, development, and yield of a crop growing on a uniform area of land under prescribed or simulated management, as well as the changes in soil water, carbon, and nitrogen over time. The minimum data needed to run these models typically include:

  1. daily weather data (max temp, min temp, precipitation, and, if possible, solar radiation);
  2. soil characteristics (soil texture, N, C, cation exchange capacity, pH, at several depths up to 50cm);
  3. crop phenology by variety/cultivar in a given environment (the timing of emergence, canopy development, anthesis [flowering], and maturity); and
  4. management practices (planting date, plant spacing, row spacing, fertilization dates), fertilizer types and amounts, harvest date, weeding and ploughing dates, ploughing and weeding implements, ploughing depth, harvested grain weight (dry), harvested biomass weight (dry).

Daily weather data can be obtained from nearby meteorological stations, many of which are shared publicly among crop modelers. Similarly, data on soil characteristics are available from the World Soil Information System (http://www.isric.org/projects/data-wosis-project) which includes the ISRIC-WISE global soil profile database (https://daac.ornl.gov/SOILS/guides/Isric.html). Crop models may include botanical parameters for well-studied varieties but further data may need to be collected for under-studied environments and varieties. The data for observed botanical traits and the data on common management practices can be collected through simple surveys for use at the household level (e.g., Mourice et al., 2014). 

Unit of analysis:

The unit of analysis is the plant biomass, grain, and stover dry weight per area of land. The accumulated nutrients, or carbon in plant biomass, can also be quantified by this method.

Limitations regarding estimating and interpreting yield from crop models

Crop modeling requires considerable investment in data collection for calibration and validation of the crop model for local conditions.

Propose Improvements