Massey Documents by Type

Permanent URI for this communityhttps://mro.massey.ac.nz/handle/10179/294

Browse

Subject: search.filters.subject.Spatial analysis (Statistics)

Search Results

Now showing 1 - 4 of 4
  • Thumbnail Image
    Item
    Predicting spatiotemporal yield variability to aid arable precision agriculture in New Zealand : a case study of maize-grain crop production in the Waikato region : a thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Agriculture and Horticulture at Massey University, Palmerston North, New Zealand
    (Massey University, 2020) Jiang, Guopeng
    Precision agriculture attempts to manage within-field spatial variability by applying suitable inputs at the appropriate time, place, and amount. To achieve this, delineation of field-specific management zones (MZs), representing significantly different yield potentials are required. To date, the effectiveness of utilising MZs in New Zealand has potentially been limited due to a lack of emphasis on the interactions between spatiotemporal factors such as soil texture, crop yield, and rainfall. To fill this research gap, this thesis aims to improve the process of delineating MZs by modelling spatiotemporal interactions between spatial crop yield and other complementary factors. Data was collected from five non-irrigated field sites in the Waikato region, based on the availability of several years of maize harvest data. To remove potential yield measurement errors and improve the accuracy of spatial interpolation for yield mapping, a customised filtering algorithm was developed. A supervised machine-learning approach for predicting spatial yield was then developed using several prediction models (stepwise multiple linear regression, feedforward neural network, CART decision tree, random forest, Cubist regression, and XGBoost). To provide insights into managing spatiotemporal yield variability, predictor importance analysis was conducted to identify important yield predictors. The spatial filtering method reduced the root mean squared errors of kriging interpolation for all available years (2014, 2015, 2017 and 2018) in a tested site, suggesting that the method developed in R programme was effective for improving the accuracy of the yield maps. For predicting spatial yield, random forest produced the highest prediction accuracies (R² = 0.08 - 0.50), followed by XGBoost (R² = 0.06 - 0.39). Temporal variables (solar radiation, growing degree days (GDD) and rainfall) were proven to be salient yield predictors. This research demonstrates the viability of these models to predict subfield spatial yield, using input data that is inexpensive and readily available to arable farms in New Zealand. The novel approach employed by this thesis may provide opportunities to improve arable farming input-use efficiency and reduce its environmental impact.
  • Thumbnail Image
    Item
    Spatial patterns in the taxonomic and dietary diversity of New Zealand rocky reef fishes : a thesis presented in partial fulfilment of the requirements for the degree of Master of Science in Biological Sciences at Massey University, Albany, New Zealand
    (Massey University, 2018) Howarth, Odette Ruth
    Latitudinal gradients of high species richness from the tropics declining towards the poles are well documented for many terrestrial and marine species. However, these broad scale patterns of numbers of species observed don’t inform as to how related these species are, or how they use food resources. By understanding taxonomic relationships between observed species I can predict how resilient these assemblages are and how environmental processes impact their distribution. Dietary diversity between species describes patterns related to mechanisms of food availability and preference of habitats or environmental niches. Marine fishes are speciose and well known taxonomically, and so comprise a useful system for studying broad-scale patterns in biodiversity. Here I examined five indices of diversity—species richness, average taxonomic distinctness, variation in taxonomic distinctness, average diet diversity and variation in diet diversity—using a historical dataset on the presence/absence of rocky reef fishes spanning most of New Zealand. I modelled these indices using boosted regression trees and mapped their distributions to the coastline at a 1km scale resolution. Additionally, I developed a new quantitative methodology to classify coastal, rocky reef fishes into homogenous diet guilds using hierarchical clustering of nine broad food items and SIMPROF multivariate analysis and modelled species richness of three of the diet guilds (herbivore, invertivore and piscivore/benthic invertivores) using boosted regression trees. This research has broadened our understanding of patterns of fish diversity, spatial patterns in diversity of diets in coastal rocky reef fishes in New Zealand. I found the indices of overall species richness, species richness of herbivores and invertivores, and average taxonomic distinctness to be highly correlated with increased wintertime sea-surface temperature indicating a latitudinal gradient to their distributions. Decreased turbidity increased average dietary diversity and species richness of the piscivore/benthic invertivore guild. Average fetch or exposure had a positive relationship with variation in diet diversity and a negative relationship with variation in taxonomic distinctness. Of secondary importance I found the indices of overall species richness, species richness of invertivores, average taxonomic distinctness and variation in diet diversity to be adversely affected by increased turbidity. Variation in taxonomic distinctness and species richness of the herbivore diet guild increased with variable and increasing salinity (respectively) while average diet diversity increased with exposure. Lastly the piscivore/benthic invertivore guild had a positive relationship with increased wintertime sea-surface temperature. Overall I found broad and fine scale environmental processes affected the species richness and taxonomic diversity of NZ reef fishes as did food availability, resource use and habitat preferences.
  • Thumbnail Image
    Item
    Geospatial threat measurement : an analysis of the threat the diatom Didymosphenia geminata poses to Canterbury, New Zealand : a thesis submitted in fulfilment of the requirements for the degree of Master of Philosophy in Geographic Information Systems in Massey University, Palmerston North
    (Massey University, 2009) Thornley, Norman John
    This thesis provides analysis of the threat Didymosphenia geminata poses to the Canterbury Conservancy of the Department of Conservation More specifically, it examines the relationship between Values, Risk and Hazard to measure the degree of threat posed by the diatom. This is the first time this type of Threat Analysis has been applied to such a problem in this region; and so will provide an important insight into the validity of the application of this methodology to an alien invasive threat. Moreover, it is the first time Values. Risk and Hazard have been modelled together to give an over all threat classification in this context. Risk mitigation is one of the variables that can be measured, managed and priced; factoring this into the model is also discussed. Qualitative and quantitative Values and Risk information is provided by Department of Conservation staff; some from their local knowledge and some from biodiversity datasets which have been collected over time. The Risk data is supplemented by fishing access data supplied by the two local Fish and Game Council Offices. Where available, further Values and Risk data has been gleaned from existing datasets in order to supplement the existing data. The Hazard data is taken from the work done by NIWA in 2005 and 2007; the latter being generated after field surveys were conducted on D. geminaia infected sites in the South Island.
  • Thumbnail Image
    Item
    Stoat trap tunnel location : GIS predictive modelling to identify the best tunnel location : a thesis submitted in fulfillment of the requirements for the degree of Master of Philosophy in Geographic Information Systems in Massey University
    (Massey University, 2009) Day, A. Mark; Day, A. Mark
    Stoats are recognised as one of the biggest threats to New Zealand's threatened species. They are difficult to control because of their biological characteristics. Currently trapping is the most common type of control technique that has a proven success rate. Research studies have shown that some traps catch more stoats than others. However the reason for this is not well documented. The effectiveness of a trap set is difficult to determine because not all trap locations are the same and not all people have the same ability to select the best location for a trap. This study uses GIS to spatially analyse stoat capture data from a control operation on Secretary Island in conjunction with commonly available vegetation, habitat, diet and home range spatial data to see if there are consistent patterns that could be used as variables in a model that would predict the best place to locate a stoat trap tunnel. The model would then be tested against a similar dataset from Resolution Island. The Department of Conservation supplied the stoat capture data from the control operations on both islands. Standard spatial analysis techniques were used to generate surfaces that combined the capture data with the vegetation, habitat, diet and home range surfaces to produce predictive surfaces. The key finding from the research was that it is possible to produce a predictive model, although one was not created because the spatial datasets were not of a high enough resolution to provide conclusive evidence that could be confidently used as a variable in a model. The spatial analysis also indicated that stoats on both islands were caught mainly in the warmer northwestern parts of the islands although the study could not determine why there was a preference for these areas. In rugged terrain like that found on both islands the location of the track network will influence where the majority of stoats will be caught.