Investigation into the bio-physical constraints on farmer turn-out-date decisions using remote sensing and meteorological data

Green, Stuart

Investigation into the bio-physical constraints on farmer turn-out-date decisions using remote sensing and meteorological data

dc.check.embargoformat	Embargo not applicable (If you have not submitted an e-thesis or do not want to request an embargo)	en
dc.check.info	Not applicable	en
dc.check.opt-out	Not applicable	en
dc.check.reason	Not applicable	en
dc.check.type	No Embargo Required
dc.contributor.advisor	Cawkwell, Fiona	en
dc.contributor.advisor	Dwyer, Edward	en
dc.contributor.author	Green, Stuart
dc.contributor.funder	Teagasc	en
dc.contributor.funder	Goddard Space Flight Center	en
dc.contributor.funder	Washington Space Grant Consortium	en
dc.date.accessioned	2019-04-24T12:07:41Z
dc.date.available	2019-04-24T12:07:41Z
dc.date.issued	2019
dc.date.submitted	2019
dc.description.abstract	Grass is the most common landcover in Ireland and covers a bigger percentage (52%) of the country than any other in Europe. Grass as fodder is Ireland’s most important crop and is the foundation of its most important indigenous industry, agriculture. Yet knowledge of its distribution, performance and yield is scant. How grass is nationally, on a farm by farm, year by year basis managed is not known. In this thesis the gaps in knowledge about grassland performance across Ireland are presented along with arguments on why these knowledge gaps should be closed. As an example the need for high spatial resolution animal stocking rate data in European temperate grassland systems is shown. The effect of high stocking density on grass management is most apparent early in the growing season, and a 250m scale characterization of early spring vegetation growth from 2003-2012, based on MODIS NDVI time series products, is constructed. The average rate of growth is determined as a simple linear model for each pixel, using only the highest quality data for the period. These decadal spring growth model coefficients, start of season cover and growth rate, are regressed against log of stocking rate (r2 19 = 0.75, p<0.001). This model stocking rate is used to create a map of grassland use intensity in Ireland, which, when tested against an independent set of stocking data, is shown to be successful with an RMSE of 0.13 Livestock Unit/ha for a range of stocking densities from 0.1 to 3.3 Livestock Unit/ha. This model provides the first validated high resolution approach to mapping stocking rates in intensively managed European grassland systems. There is a demonstrated a need for a system to estimate current growing conditions. Using the spring growth model constructed for estimating stocking density a new style of grass growth progress anomaly map in the time-domain was developed. Using the developed satellite dataset 1 and 12 years of ground climate station data in Ireland, NDVI was modelled against time as a proxy for grass growth This model is the reference for estimating current seasonal progress of grass growth against a ten year average. The model is developed to estimate Seasonal Progress Anomalies in the Time domain (SPAT), giving a result in terms of “days behind” and “days ahead” of the norm. SPAT estimates for 2012 and 2013 are compared to ground based estimates from 30 climate stations and have a correlation coefficient of 0.897 and RMSE of 15days. The method can successfully map current grass growth trends compared to the average and present this information to the farmer in simple everyday language. This is understood by the author to be the first validated growth anomaly service, and the first for intensive European grasslands. The decisions on when to turn out cattle (the turn out date (TOD)) from winter housing to spring grazing is an important one on Irish dairy farms which has significant impacts on operating costs on the farm. To examine the relationship of TOD to conditions, the National Farm Survey (NFS) of Ireland database was geocoded and the data on turn out dates from 199 farms across Ireland over five years was used. A fixed effects linear panel data model was employed to explore the association between TOD and conditions, as it allows for unobserved variation between farmers to be ignored in favour of modelling the variance year on year. The environmental variables used in the analysis account for 38% of the variance in the turn out dates on farms nationwide. National seasonal conditions dominate over local variation, and for every week earlier grass grows in spring, farmers gain 3.7 days in grazing season but ignore 3.3 days of growth that could have been used. Every 100mm extra rain in spring means TOD is a day later and every dry day leads to turn out being half a day earlier. A well-drained soil makes TOD 2.5 days earlier compared to a poorly drained soil and TOD gets a day later for every 16km north form the south coast. This work demonstrates that precision agriculture 1 driven by optical and radar satellite data is closer to being a reality in Europe driven by enormous amounts of free imagery from NASA and the ESA Sentinel programs coupled with open source meteorological data and models and new developments in data analytics.	en
dc.description.status	Not peer reviewed	en
dc.description.version	Accepted Version
dc.format.mimetype	application/pdf	en
dc.identifier.citation	Green, S. 2019. Investigation into the bio-physical constraints on farmer turn-out-date decisions using remote sensing and meteorological data. PhD Thesis, University College Cork.	en
dc.identifier.endpage	192	en
dc.identifier.uri	https://hdl.handle.net/10468/7795
dc.language.iso	en	en
dc.publisher	University College Cork	en
dc.rights	© 2019, Stuart Green.	en
dc.rights.uri	http://creativecommons.org/licenses/by-nc-nd/3.0/	en
dc.subject	Remote sensing	en
dc.subject	Grassland	en
dc.subject	Panel analysis	en
dc.thesis.opt-out	false
dc.title	Investigation into the bio-physical constraints on farmer turn-out-date decisions using remote sensing and meteorological data	en
dc.type	Doctoral thesis	en
dc.type.qualificationlevel	Doctoral	en
dc.type.qualificationname	PhD	en
ucc.workflow.supervisor	f.cawkwell@ucc.ie