"I can’t be green if I’m in the red": combining precision agriculture and remote sensing technologies for sub field and regional decision making.

Abstract

Balancing sustainable agricultural production with environmental, social, cultural and community objectives under the uncertainty of the impacts of climate change on rural livelihoods has become an increasing priority worldwide. This may mean land-use pattern, that have evolved over the last decades may be suboptimal. Environmental degradation but also economic opportunities for climate change mitigation from carbon sequestration may support alternative land-use scenarios. However, the majority of cost of such changes is expected to be borne by the landholder and adoption of alternative land uses will only occur if profit from traditional cropping practices is comparative to new options, namely, in areas where the economic opportunity cost is low. Precision agriculture has shown that yield variation in fields can be substantial and here lies the potential that is explored in this thesis. Precision agriculture provides data with a spatial resolution that is fine enough to reflect the spatial variably within fields. If unproductive patches can be allocated to more environmentally friendly use, both the environment and farm economy may benefit. However, inherent problems exist with the technology and these need to be addressed before the information can be used in the decision making process. A preparation step in this thesis is therefore to evaluate a suite of targeted algorithms to remove a substantial amount of yield mapping errors. This thesis examines the degree of spatial and temporal variability and estimates a potential range of economic opportunity costs that might be associated with reallocation of land to different use. Although dependent on the interplay between the spatial and temporal variability of yield and the price volatility of international commodity markets, a likely scenario shows that about 50% of the land may be taken out of production with only a 25% reduction in income. Regional land managers do not have access to precision agriculture data because yield mapping data does not exist at a scale or temporal dimension required for regional analyses. This thesis shows that it is possible to creating high resolution estimates of economic performance at a broad scale by extrapolating yield mapping data from early adopters to an entire study area using remotely sensed imagery over numerous seasons. This also has strong benefits for landholders who do not have long time series of yield data. By using satellite remote sensing they may be able to leap frog the long phase of yield map archiving giving them the ability to make management and land use decisions sooner. This thesis suggests that high resolution yield estimates combined with financial estimates of production can identify cropping areas with marginal income returns. This type of information may facilitate adoption of a mix of environmentally friendly land uses in the cropping landscape without significant financial repercussions to the grower. Additionally, the mapping of this information will act as a critical sounding board between the land holder and the catchment manager where conflicting objectives of economic and environmental outcomes can be compared.