Small plots and research scales are not as realistic as productivity from field plots when making economic assumptions in any project. We detected misleading information and sampling methods on energy crops and show why small plots are not the right path to follow.

When all this about energy crops started, most scientific information and accumulated knowledge on “perennial grasses” and woody plantations came from the “forage and forest sciences”. It was 30 years ago more or less when lignocellulosic energy plantations began to be studied systematically for energy uses and biomass. In forage sciences most measures considered always grazing management and most evaluations have to be reconsidered for large scale projects with grass cuttings as in your case. In woody plantations, several experiences considered large rotations and wood quality that changed in short rotation coppice techniques for biomass chips production. Several studies have been taking place in the last decade with more than 40 species. Plant densities, fertilization trials, different crops producing herbaceous and woody biomass, studies on photosynthesis efficiencies, etc. Many of  these reports, have  different levels and scales (from “leaf” level, to plant, crop, canopy, and farm but also at regional scales). However to know productivity levels for a certain condition is always the “KEY” issue on bioenergy crops assessments.

Our company can save time and money providing reliable information on energy crops performance as well as realistic estimations with several species.

Why microplots cannot provide  realistic yield estimations?

Regarding controlled conditions and  small plots, there is a lot of information in experimental sciences and applied methodologies. Most are focused on  leaf, plant and canopy levels. Such studies are always needed and can be considered but would not be a safe way to establish your productivity levels. In particular, many species like Giant King grass but also other strains of pennisetum purpureum, arundo donax and miscanthus, can provide huge expectations in most tropical environments. Spectacular potential yields and radiation interception capacity in ideal conditions is feasible, as well as high water and nitrogen use efficiencies. We will provide some analysis about that buy again, don’t think that yields measured are possible to  compare or extrapolate to a large commercial plantation.

Establishing commercial yields using only small plots is dangerous. In small plots, a large percentage  of plants and biomass accumulated on the ground surface (aerial biomass) have contact with empty spaces that are plenty of resources (light, water, nutrients).  Those resources are always available and plants in the permiter can use them. In large scale farms, that does not happen as the individual plants in the permiter represent only a very small part of the total biomass production. All plants inside the parcel will have competition for resources and limitations to productivity.

Figure 1: Small plots trials and border effects. Photo: C3 perennial grass.

The European experience implementing programs with energy crops in large scale demonstration

Poland, Hungary, Italy, Portugal, Spain, England, Germany, Greece and Sweden have been involved in several experimental and demonstration programs during last decade. In most of them millions of euros have been used for research, large multi-crop and multi-region scales, planting material developments and nurseries  or seed breeding and  specific machinery manufacture for new energy crops and management for biomass production. We have been participating on the experimental and development and commercial implementation with woody and herbaceous crops in most of those programs with more than 35 experts in our network. Most of their results have demonstrated that extrapolation and upscaling results  from experiments, trials and small plots is not realistic or advisable. We have been involved in those programs and use that information as a base.

The estimations using small plots often consider border effects (see Figure 1 above). A typical case we observe every day is reflected in a study we did years ago with several research institutes in Europe (subhumid-semiarid regions of Spain in this case, see Figure 2). Experimental plots are often used to estimate yields and compare genotypes (varieties, species, subspecies)  and management techniques. The crop yields are great in semiarid environments but when we passed to large scale projects, we found that ranges of biomass productivity were 4 times lower (about 20 wet tons /acre with 2 cuttings). This is a winter cool grass and it is not comparable with tropical grasses. It is just an example to show you how small plots create misleading conclusions.

Now, in Figure 2 and Figure 3 you can see a typical grass strip we do in experimental trials. We call it MACROPARCEL. We developed them since we needed to demonstrate commercial equipment and see realistic yields and weights. Grass strips are often sized 1/2 or 1/3 of acre per treatment, sometimes 3-5 repetitions for each variety or fertilizer/density trial.

Grass strips in European research programs to evaluate commercial yields. Photo: C3 drought tolerant energy grasses in temperate areas (2013).

You need to know, that neither your planting material supplier or an excellent researcher will give you enough information from its own plantations in large scale projects. They will give information on small plots, trials and nurseries where they have been involved during last few years. Additionally,  those yields often have biased information and do not consider the lifetime of the plantation that is often reduced compared to the lifetime of a small plot.

Small plots often have biased results from sampling methods with no replicates or repetitions. In case they have included enough repetitions to sample variability, nobody will know standard deviations and variability or certainty levels for claims.  

Most experiences with energy crops in tropical areas are in this case. They are often experiences coming from  ideal conditions in small areas (where water, soil and other factors are always well managed and controlled).  But we all know that farming in large scale projects (hundreds / thousands  of acres ) implies to cultivate on heterogeneous soil fertility scenarios and may certainly have a lot of environment x genotype interactions.

Reliable data source and productivity estimations:

If you have tried small plots, your data are a promising result. However you won’t have shown yourself or your investors enough elements to establish average yields in dry matter per year. You will need other methods to make right assumptions for  large plantations or expected average situation to be assumed with certainty.

Grass strips and trials with Pennisetum purpureum (drying field treatments to reduce transportation costs). Test pots size: 6000 sq ft. per repetition (5 repetitions).

Our company and experts have been involved in research and consultancies in the last 20 years. We often use small plots and compare small VS large plots. Small plots frequently produce higher yields (from 50 to 200 % higher) compared to large parcels (that should be above 1/3 acre or more as minimum size per each replicate). We often make strips o 1/3 acre are 3 to 5 repetitions in different areas of the farm. That is much more trustable but the number of repetitions and plot size need to consider specific variability and soil conditions or even slopes. And it is what any seed supplier  company does when establishing trials (see Monsanto, Pioneer, etc.). If small plots are going to be used, we can also suggest a trial design for a experimental program.

Some scientific  literature supporting this document:

1. A comparison of two methods to predict the landscape-scale variation of crop yield Original Research Article.  Soil and Tillage Research, Volume 58, Issues 3–4, March 2001, Pages 163-181-  F.C Stevenson, J.D Knight, O Wendroth, C van Kessel, D.R Nielsen
2. Field plots for the practical estimation of potential yield Original Research Article. Scientia Horticulturae, Volume 10, Issue 4, June 1979, Pages 309-316 R. Thompson, H. Taylor
3. Estimation of Crop Yield Distribution: Implication for Crop Engineering Risk Original Research Article. Systems Engineering Procedia, Volume 3, 2012, Pages 132-138. Wang Xiang, Li Yunxian, Qian Zhenwei, Shen Zeliang
4. Influence of images recording height and crop growth stage on leaf cover estimates and their performance in yield prediction models Original Research Article. Crop Protection, Volume 18, Issue 8, September 1999, Pages 501-508- M. Ngouajio, G.D. Leroux, C. Lemieux
5. Spatial variability of switchgrass (Panicum virgatum L.) yield as related to soil parameters in a small field. Di Virgilio, Nicola; Monti, Andrea; Venturi, Gianpietro.  Field Crops Research , Volume 101 (2). Elsevier. Mar 5, 2007
6. The discrepancy between plot and field yields: Harvest and storage losses of switchgrass. A. Monti*, S. Fazio, G. Venturi. Department of Agro-environmental Science and Technologies, University of Bologna, viale Fanin 44, 40127 Bologna, Italy.
7. Spatial variability of Switchgrass (Panicum virgatum L.) yield as related to soil parameters in a small field. Nicola Di Virgilio, A. Monti, G. Venturi. Institute of Biometeorology IBIMET-CNR, Via P.Gobetti 101 I, 40129 Bologna, Italy. Field Crops Research 101 (2007) 232–239. biomass and bioenergy 33 (2009) 84 1–847