Seminar on Innovations in Agri-Tech

5th May 2016; 14.00 – 18.00

The Sainsbury Laboratory, Cambridge University, Bateman Street, Cambridge, CB2 1LR

Links:

Centre for Global Equality

CambPlants Hub

University Strategic Research Initiative on Global Food Security (GFS)

CropPerformance Ltd,

Department of Plant Sciences

Presentation 1: Abstract [Click here for full paper]

Sustainable Perennial Crop Production in the Tropics: Understanding Yield Variation Under a Changing Climate

Stephanie Race D.MarcusBellett-Travers, Crop-Perfromance Ltd, Cambridge

In tea production there is a need to be able to monitor and plan productivity. Essential to this is knowledge of the impact and implications of crop nutrition, crop water status and management practices on the growth and development of the crop and the resulting yield. This understanding not only relates to yield maximisation but also the optimisation of cropping efficiency and management productivity. This in-turn leads to profitability. Crop Performance Ltd. engaged in a collaborative project to create a yield model that has the potential to predict monthly yield data for made tea at the field level. The model was developed using a hind casting approach leveraging 15 years of historical NDVI data from satellites and historical meteorology data to determine the capacity for light interception by the tea canopy. This is then combined with ground collected solar radiation data to determine a potential for photosynthetic output by the crop. This potential is then converted to an estimate of actual dry matter accumulation by taking account of the effects of critical environmental factors: temperature, evaporative demand, soil- water-status, crop-water-status, nutrient-status of the crop and nutrient uptake by the crop. This accumulation of dry matter is then converted to a yield of made-tea by proportioning the dry matter to various parts of the crop: the absorbing roots; structural (woody) parts of the plant above and below ground; leaves; and those leaves that are harvested (that become the yield of made tea). Critical to the workings of this model are the collection of robust environmental data that is pertinent to each estate. The outcomes inform a yield forecast, water and nutrient status of the crop.

The crops are grown on Kenyan estates are typical of tea grown in East Africa and are from either clonal material or form seedlings (seedling tea) of mixed genetic origin (Tea Research Foundation, 1990). As a woody perennial plant, the crop develops a woody architecture of both shoots and roots from which the absorbing roots and leaves grow respectively (Pallardy, 2008). Although the absorbing roots and leaves are often seen as more biologically active material; the vascular system of the woody material plays an important part in the physiology of the plant as the conduit for water, sugars and nutrients up and down the stem as well as the production of phytohormones and other important metabolic chemicals. In addition growth is not just a the terminal shoot and root buds but also in the cambium of woody tissues which result in growth rings (Pallardy, 2008).

The amount of amount of radial growth in woody tissues is similar each year for similar environmental conditions and is the reasoning behind using tree growth ring data to study historic weather conditions and climate change (Pallardy, 2008). This means that the partitioning of dry matter is heavily influenced by the amount of differentiation in woody stems (the degree of branching) and the length of each element (branch). The distribution of dry matter between woody tissues and leaves is one of the major causes of variation in yield both between clones and with age (Burgess and Carr, 1996; Ng'etich and Stephens, 2001; De Costa et al, 2007; Dutta, 2011).

Crop management includes a number of practices that alter growth and development (Tea Research Foundation, 1990; Dutta, 2011). Harvesting leaves by plucking, by its nature alters the canopy and removes canopy dry matter and it is anticipated that this will have a subtle impact on canopy cover. Pruning is typically carried out in four year cycles and is a more aggressive removal of structural shoot material. Following both plucking and pruning, not all leaf material is removed and there is a pool of retained leaves (leaves remaining after plucking or pruning) that continue to photosynthesise and contribute dry matter to both the growth and development of the plant and to the next harvest (De Costa et al, 2007). The partitioning of dry matter between retained and harvested leaves also varies with age but there is less apparent variation between clones (Burgess and Carr, 1996).

While light interception by the crop is the important driver of dry matter production (Ng'etich and Stephens, 2001), the production, development and promotion of yield is as much about dry matter partitioning as it is about dry matter production (Burgess and Carr, 1996; De Costa et al, 2007; Dutta, 2011). The crop intercepts light with leaves and uses it to make carbohydrates via photosynthesis. Traditionally crop models estimate light interception by the crop via measurements of incident radiation or PAR and a description of the light profile down the canopy from LAI measurements and a corresponding extinction coefficient (Pallardy, 2008). However, LAI is a two-dimensional description of a three-dimensional architecture and is prone to error. The introduction of NDVI and the possibility of its remote measurement via satellite or drone, allows a more realistic description of light interception with respect to how the canopy actually functions.

NDVI is the ratio of reflected near infra-red radiation (NIR) minus reflected red-light in the visible spectrum (VISR) to the NIR plus VISR. The absorption of light by the chlorophyll pigments in the visible range and the strong reflectance in the NIR range means that the more leaves and chlorophyll a plant has, the greater the degree of VISR absorption and the greater the degree of NIR reflectance such that the ratio approaches 1. Whereas when plants grow with poor pigment contents and few leaves or loose pigments and leaves under stress, then the ratio approaches 0 (Ustinet al., 2009). The NDVI bears a near linear relationship with LAI (up until NDVI equals 1), the fraction of photosynthetically active radiation (fPAR) (Mynemi and Williams, 1994) and photosynthetic biochemical content. By using NDVI it is possible to avoid using LAI and extinction coefficients and simply correlate NDVI to dry matter production via measurements of incident radiation. Generally, vegetation indices and normalised ratios have been shown to be a strong indicator of photosynthetically active biomass (Van Der Meer et al., 2001) and avoid the errors associated with LAI.

The water use efficiency of tea crops has been the focus of many studies, with rooting depth and root distribution identified as traits of interest (De Costa et al, 2007). The ability of roots to penetrate the soil is also important, with impedance sited as a barrier to root depth (Pallardy, 2009). The soils of the Unilever estates are sandy loams to sandy clays typically consisting of 60%Sand:16%Silt:24%Clay (Data supplied by Unilever). Without the presence of organic matter the soil available water capacity would only be around 130 mm.m-1 (Marshall et al, 1996). Therefore the presence of soil organic matter will have a significant impact on the yield of the cropping system when rainfall is limited.

Nutrients are applied to the crop as inorganic fertilisers commonly containing a ratio of 25:5:5 Nitrogen (N), Phosphorus (P) and Potassium (K) with an addition of Sulphur (S) occasional Magnesium (Mg) and Zinc (Zn) area also added (Data supplied by Unilever). The organic matter will also contribute or deplete the nitrogen content of the soil depending on the soil conditions. Turnover of high carbon containing compounds such as cellulose and lignin that have a low nitrogen content will tend to remove soluble nitrogen from the soil. When soil is saturated for long periods denitrification will also remove soluble soil N. In contrast when soil is well aerated nitrogen fixing bacteria will add soluble N to the soil. Well stabilised organic matter will also contribute a slow release of soluble N (Marschner and Rengel, 2011).

The uptake and distribution of nutrients within the crop must be aligned to the production and partitioning of dry matter, otherwise leave, absorbing roots or the cambium will become deficient and their physiological processes optimised (Römheld, 2011).

To model the productivity of the crop, light interception and conversion to dry matter is the starting point in the process of yield development but is optimised by the presence of water and nutrients in leaves in optimum amounts. Equally the uptake and distribution of water and nutrients throughout the crop relies on the health of absorbing roots and the vascular system. This intrinsically links growth to the water cycle and nutrient cycles. The Partitioning determines the proportion of dry matter that becomes yield but also drives light interception, water uptake and nutrient uptake. The influences of management practices and harvest methods will change and control the partitioning of dry matter and affect yield.

Presentation 2: Abstract [Click here for Full paper]

Biotechnology to control disease in African crops

David Baulcombe, Reguius professor of Botany, Plant Sciences Dept, Cambridge University

Molecular biology has provided a good understanding of how disease resistance operates in plants and it points to novel interventions that could be used in many settings including Africa. Some of the most promising approaches are based on a novel type of crop protection chemistry in which the aim is to apply chemicals that activate the host defense systems rather than being microbicides targeted to pests and pathogens. Some of these novel compounds can be applied to seeds and have long lasting effects so that problems with application are avoided.

Other approaches involve the use of genetics. With molecular markers we can enhance conventional breeding for disease resistant varieties or use GM. In my laboratory we are developing a GM approach to tackle maize lethal necrosis disease – a problem for farmers in East and Central Africa. The approach exploits an RNA-based mechanism and so it does not require the crop to produce any novel proteins. Our current project is to prove the concept although, with a suitable regulatory framework for GMOs in place, the output could be directly useful in the field. In a next generation variation on this approach we could use gene editing so that there are no transgenes involved.

In my talk I will discuss the benefits and disadvantages of these various approaches.

Panel Discussion

Panellists: Nigel Melican, Will Simonson, Stephanie Race, David Baulcombe

Facilitator: Keith Virgo

Name / Question/ Response
Keith / What sort of platforms are you using for remote sensing in your case study in Kenya?
Stephanie / For the case study that we presented, this was a model development exercise so we are doing hind-casting, using a LANDSAT archive which is satellite remote sensing platform. But I think what is more pertinent has to do with what type of spectroscopy needs to be provisioned for what particular end user application. So it is not just trying to get the same measurement but on a different platform, it has to do with measurement and spatial scale. Some of the things that David Baulcombe presented on disease resistance, in the process of introducing new disease resistance varieties,is something that happens between the time that breeding happens and the varieties get introduced to growers.One of things that we are involved in using spectroscopy for is in field phenotypes where we are looking at the performance of those varieties and how traits actually interact with the environment before they get introduced to growers. So that we can sort of prove it out. So that is an example of doing in-field phenotyping.
Keith / In the latest Agriculture for Development journal, there is an article by Geoff Hawtinon the CGIAR organisations, who are using drones now. One is phenotyping to pick up different types of problems, that they can then do analysis on, breed from and identify the difference. The other very interesting one I thought was that in research plots, you can have drones overhead and they can report back if a herd of cattle break into the research trial and you can get them out before they do any damage, A very practical use.
Audience / The four different remote sensing methods that you have, how did you combine the data and how did you reconcile/ coordinate systems and temper resolution?
Stephanie / There are two different concepts. One is different platforms. Essentially different platforms that are provisioning dataat different spatial coverages, meaning what geography you are actually monitoring, and how frequently. The other is the cameras so they are provisioned on the platforms which have different capabilities based upon the image resolution, the data, and the type of spectroscopy, meaning that data that you are actually picking up and what you are doing with that light interception data. So an example would be measuring something like NVDI using a spectral camera, and you can provision that off a drone or off a satellite. You can do both at the same time.
If you have a cloudy day in your time series you could provision the cloudy days from the UAV (drone). In clear days, from satellite, then you end up with one time series where you have multi spectral data that is provisioned from two different platforms on one piece of land over a period of 20 weeks or so.
On tea estates you have shade trees anda whole range of agri-forestry features. One of the things we do with projects is just describing this hind-casting approach.So we are relying on an archive. Mostly archives that exist today are on satellites, but we have archives from air photography that we have run on specific geography,like one in California that we work with, which has been run on a daily time step since 2009.
Back to answering your question, there is a way to actually leave the data together on a mosaic but dealing with sometimes different spatial coverage you may have different angles. There are all of these variables to address.It is a complicated answer. We are developing something, thathasbeen called the vector,a technical term for how we address the image analysis and the intractability issues with multiple cameras provisioned from multiple platforms on a single piece of land for one time series of acrop, for example, in one single year.
Keith / Genes that you developfor plants,who owns them?
David / It depends. For work that is done in the public sector, they will often be owned by either the institutions or by charities who funded the work. The motivation of those types of organisations is often to ensure that the technology is used as widely and as appropriately as possible. There are examples of disease resistance genes which have been isolated by companies, and of course they are then owned by the companies. With our African Maize Lethal Necrosis project, which we are doing under funding by BBSRC, I think the intellectual property situation would be so complicated that there will be no ownership of that at all. It is certainly my intention that we would make surethat the germplasmwould be available to whoever wanted to use it.
Keith / It is the principle of bio-technology, anyway. It is supposed to be freely available?
David / In some people’s eyes, yes, and certainly that would be my aspiration. We have a joint project between my department and the John Innes Centre. It is a synthetic biology project called ‘Open Plant’. Part of the ethos of ‘Open Plant’ is to develop bio-technology comparable to the open source approach that has been used in the software industry, where it has actually promoted rather than restricted innovations. In my view it is that the intellectual property approach in certainly plant bio-technologies until now has probably inhibited, rather than stimulated, innovations.
Audience / Overly wealth-orientated entities are filing patents all over the shop, particularly in territories across the world where we should not be asserting intellectual property rights, rather we should be giving it away. Having worked in the IP space for a long time, I don’t think I feel too guilty in patenting in America and across Europe but certainly, however, the idea of patent across Sub-Saharan Africa strikes me as evil and wicked.
Audience / In the US, you can’t patent just isolated genes that are heavily modified in any way. Situations are different in Europe, you can get patent protection for isolated genes. It varies around the world basically, in terms of how the case law has developed. In order to get patent protection for something, you need to prove that what you have done is novel and not obvious as well. It is not just a case of an exception to prove that it was particularly difficult and not obvious to isolate particular genomes with some surprising advantages. It used to be a lot easier when sequences were usually identified but now, with genome sequencing, it is usually not difficult to find the gene.