Share

Production Team explores yield implications of bringing new land into production

Tags:

Posted: November 16, 2015

One way to increase a region’s capacity to meet its food needs is to bring new land into production; another is to change the mix of crops produced on existing farmland. But what are the potential yields we could expect from new or converted land? That’s the question behind a new tool developed by the Production Team — a productivity index that will help quantify the production capacity of all the arable land in the Northeast.
Graph showing cumulative potential of New York's farmland. See bottom of page for full-size image and caption.

Graph showing cumulative potential of New York's farmland. See bottom of page for full-size image and caption.

“When we talk about bringing new land into production, we could assume that the next acre brought into production is going to be just as productive as the last acre. Similarly, if we looked at converting pasture into vegetable production, for example, we could assume that it will be just as productive for growing vegetables as it was for growing forage,” said Christian Peters, a Tufts University assistant professor and co-leader of the Production Team. “The purpose of the productivity index is to actually test these assumptions, or at least to use data to inform what the relationship is between a given land type and its potential productivity for different crops.”

Peters and his colleagues are combining data from several sources to develop the index. They started with the USDA’s Cropland Data Layer, which uses satellite imagery to depict at a 30-meter scale how land is used and what crops are being grown on it. Using these geospatial data, the team assembled eight categories of land use, organized by crop needs and growth cycles: vegetables and melons, berries and tree crops, other cultivated land and alfalfa, hay and pasture, fallow and idle cropland, shrub and scrubland, forest, and all other land generally considered unsuitable for farming, like wetlands.

Next, the team worked with gSSURGO, a database from the National Resources Conservation Service that provides detailed soil geographic information that can be combined readily with map layers from other datasets. Using the tool’s land-capability classifications, which rank the suitability of soils in the continental US for crop cultivation, the team screened out land from their dataset that isn’t suitable for growing crops. They also made use of another gSSURGO asset — the National Commodity Crop and Productivity Index (NCCPI) — which, on a scale from zero to one, ranks soils on their inherent ability to produce commodity crops without irrigation. Using these tools, the researchers can determine the arability and expected productivity of any given location in the US.

Taken together, the eight land-cover categories gleaned from the Cropland Data Layer and the relative productivity data from gSSURGO form the basis of the team’s productivity index, said Ashley McCarthy, a Tufts University doctoral candidate who is working with Peters on the index.

“For every piece of land we have, we know which category it belongs to. We also have the NCCPI value that tells us relatively how well you could produce crops on it,” said McCarthy. “From there, we can estimate the mean and median potential yields we could expect for each of the eight land-cover classes we identified.”

For example, using corn for their preliminary calculations, the researchers have estimated the potential yields that could be expected in New York State, if all the eight classes of land in New York were planted to corn (see graph). They hope that as they repeat this calculation for each state, they will be able to determine a relationship between average potential yields and land type, and how that relationship may vary geographically.

They also intend to expand their calculations to include non-commodity crops. But first, they must translate the NCCPI rating of relative productivity — which is only provided for commodity crops like corn, soy, and wheat — into an estimate of absolute productivity for other crops, like fruits and vegetables. This isn’t a straightforward calculation, though, in part because soil-specific estimates of yield for most crops simply aren’t available.

“At best you might find yield data for fruits and vegetables by state, and in some cases you might find sub-state data, but often you don’t. There just isn’t enough data collection for the National Agricultural Statistical Service (NASS) to report estimates at that level,” said Peters. However, Peters and colleagues plan to summarize data by “farm resource regions” — groups of counties that are generally similar to one another agriculturally. Peters explained that these data allow him to estimate yields for counties in states where data on specific crops is lacking.

Using these actual yield data along with their estimates of relative productivity, the researchers will calculate actual yield estimates for a number of crops for each of the eight land-cover classes.

Establishing this productivity index will not only help the Production Team in its effort to quantify the region’s production capacity, but also will assist Distribution Team members with their fresh-produce models that explore how various supply-chain scenarios might prompt a shift in the locations of fresh-produce production.

“If, for example, consumers wanted substantially more fresh produce, our model examines whether the current growing locations are best suited for supplying the new demand, or might other regions be better suited to meet more of the demand,” said Patrick Canning, an economist with the USDA Economic Research Service and a member of the Distribution Team. “To answer this, we need to know where suitable land exists to grow produce and how much product can this land supply. The efforts by the Production Team to develop region-specific soil productivity indexes is ideally suited to our modeling purposes.”

More information about the Production Team’s activities are available here.

-- by Kristen Devlin


A graph depicting the productivity of NY state arable land if all land was planted to corn. Non-linear slope, with dots indicating different land classes.

Caption: The researchers have estimated the potential corn yields (vertical axis) that could be expected in New York State, if all the eight classes of land in New York were planted to corn. On the horizontal axis, the land classes are represented by dots in order of their production potential, starting with the most productive land class on the left and progressing to less productive land classes. These classes, from left to right, are: vegetables and melons, berries and tree crops, other cultivated land and alfalfa, hay and pasture, fallow and idle cropland, shrub and scrubland, forest, and all other land generally considered unsuitable for farming, like wetlands. As acres from less productive land classes are brought into production, the average yields from those acres decline, as demonstrated by the decrease in slope of the line.