Farm-Scale Crop Yield Prediction from Multi-Temporal Data Using Deep Hybrid Neural Networks

Farm-scale crop yield prediction is a natural development of sustainable agriculture, producing a rich amount of food without depleting and polluting environmental resources. Recent studies on crop yield production are limited to regional-scale predictions. The regional-scale crop yield predictions usually face challenges in capturing local yield variations based on farm management decisions and the condition of the field. For this research, we identified the need to create a large and reusable farm-scale crop yield production dataset, which could provide precise farm-scale ground-truth prediction targets. Therefore, we utilise multi-temporal data, such as Sentinel-2 satellite images, weather data, farm data, grain delivery data, and cadastre-specific data. We introduce a deep hybrid neural network model to train this multi-temporal data. This model combines the features of convolutional layers and recurrent neural networks to predict farm-scale crop yield production across Norway. The proposed model could efficiently make the target predictions with the mean absolute error of 76 kg per 1000 m2. In conclusion, the reusable farm-scale multi-temporal crop yield dataset and the proposed novel model could meet the actual requirements for the prediction targets in this paper, providing further valuable insights for the research community.

Evaluating the sensitivity of the LUCI model to GIS datasets to enable robust farm management decisions in New Zealand

<p>Degradation of water quality is a major issue in New Zealand, to which the loss of nitrogen, phosphorus and sediment from agriculture into waterways contributes significantly. To predict and manage diffuse pollution from intensive agriculture it is vital that models are able to spatially map the sources, flows and sinks of nutrients in the landscape and spatially target mitigations. This study investigates the application of one such model, the Land Utilisation Capability Indicator (LUCI). Used in conjunction with OVERSEER, LUCI is a powerful tool to support farm scale land management decision-making. LUCI includes soil, topography and landcover datasets in its analysis. This thesis examines how the quality and resolution of each dataset affects LUCI’s output. Six different case studies are examined, across a range of New Zealand farming systems. This is the most comprehensive study, to date, of LUCI’s sensitivity to input datasets. The results suggest that LUCI nutrient loading estimates are primarily sensitive to soil order, and therefore to changes in order classifications. Utilising different soil datasets in the LUCI model resulted in varying nutrient load predictions. This sensitivity is primarily attributed to the differing hydraulic and phosphorus retention capabilities of the respective soil orders. To test the sensitivity of LUCI to digital elevation model (DEM) resolution, multiple DEMs with varying spatial and vertical resolution were tested. These results strongly indicate that particularly fine resolution DEMs are required to accurately model flat landscapes. It was recognised that LUCI was not using all of the relevant data available in Landcare Research’s S-Map database. LUCI was modified to use more of this information, and alternative methods of incorporating sibling level data in both LUCI and OVERSEER were investigated. Finally, avenues for future development are suggested. Overall, this thesis highlights the potential LUCI has to play a key role in farm scale environmental management.</p>

Evaluating the sensitivity of the LUCI model to GIS datasets to enable robust farm management decisions in New Zealand

<p>Degradation of water quality is a major issue in New Zealand, to which the loss of nitrogen, phosphorus and sediment from agriculture into waterways contributes significantly. To predict and manage diffuse pollution from intensive agriculture it is vital that models are able to spatially map the sources, flows and sinks of nutrients in the landscape and spatially target mitigations. This study investigates the application of one such model, the Land Utilisation Capability Indicator (LUCI). Used in conjunction with OVERSEER, LUCI is a powerful tool to support farm scale land management decision-making. LUCI includes soil, topography and landcover datasets in its analysis. This thesis examines how the quality and resolution of each dataset affects LUCI’s output. Six different case studies are examined, across a range of New Zealand farming systems. This is the most comprehensive study, to date, of LUCI’s sensitivity to input datasets. The results suggest that LUCI nutrient loading estimates are primarily sensitive to soil order, and therefore to changes in order classifications. Utilising different soil datasets in the LUCI model resulted in varying nutrient load predictions. This sensitivity is primarily attributed to the differing hydraulic and phosphorus retention capabilities of the respective soil orders. To test the sensitivity of LUCI to digital elevation model (DEM) resolution, multiple DEMs with varying spatial and vertical resolution were tested. These results strongly indicate that particularly fine resolution DEMs are required to accurately model flat landscapes. It was recognised that LUCI was not using all of the relevant data available in Landcare Research’s S-Map database. LUCI was modified to use more of this information, and alternative methods of incorporating sibling level data in both LUCI and OVERSEER were investigated. Finally, avenues for future development are suggested. Overall, this thesis highlights the potential LUCI has to play a key role in farm scale environmental management.</p>

The Sustainable Treatment Effect of Constructed Wetland for the Aquaculture Effluents from Blunt Snout Bream (Megalobrama amblycephala) Farm

In aquaculture, constructed wetland (CW) has recently attracted attention for use in effluent purification due to its low running costs, high efficiency and convenient operation,. However, less data are available regarding the long-term efficiency of farm-scale CW for cleaning effluents from inland freshwater fish farms. This study investigated the effectiveness of CW for the removal of nutrients, organic matter, phytoplankton, heavy metals and microbial contaminants in effluents from a blunt snout bream (Megalobrama amblycephala) farm during 2013–2018. In the study, we built a farm-scale vertical subsurface flow CW which connected with a fish pond, and its performance was evaluated during the later stage of fish farming. The results show that CW improved the water quality of the fish culture substantially. This system was effective in the removal of nutrients, with a removal rate of 21.43–47.19% for total phosphorus (TP), 17.66–53.54% for total nitrogen (TN), 32.85–53.36% for NH4+-N, 33.01–53.28% NH3-N, 30.32–56.01% for NO3−-N and 42.75–63.85% for NO2−-N. Meanwhile, the chlorophyll a (Chla) concentration was significantly reduced when the farming water flowed through the CW, with a 49.69–62.01% reduction during 2013–2018. However, the CW system only had a modest effect on the chemical oxygen demand (COD) in the aquaculture effluents. Furthermore, concentrations of copper (Cu) and lead (Pb) were reduced by 39.85% and 55.91%, respectively. A microbial contaminants test showed that the counts of total coliform (TC) and fecal coliform (FC) were reduced by 55.93% and 48.35%, respectively. In addition, the fish in the CW-connected pond showed better growth performance than those in the control pond. These results indicate that CW can effectively reduce the loads of nutrients, phytoplankton, metals, and microbial contaminants in effluents, and improve the water quality of fish ponds. Therefore, the application of CW in intensive fish culture systems may provide an advantageous alternative for achieving environmental sustainability.

Factors Related to Profitability of Agritourism in the United States: Results from a National Survey of Operators

Agritourism is a growing area of the tourism sector with many positive social and economic benefits for farmers, their communities, and for tourists. While researchers have been studying the phenomenon for several decades, factors that lead to profitable outcomes for agritourism operators are still not well understood, hindering the effectiveness of agritourism development and the systems of support available to farmers. Using a survey of 1834 farms and ranches open to visitors in the United States, the goal of this study is to identify the factors that influence the profitability of agritourism operations. This study shows that several factors have positive associations with increased agritourism profitability, such as the number of years of experience of the operator, farm scale (acreage and total farm revenue), providing on-farm product sales, and offering events and entertainment. Off-farm product sales and being a female operator have a negative association with profitability in agritourism. We discuss the implications of our findings on agritourism operators, suggest their utility for tourism planning and rural community development professionals, and offer suggestions for future research.

Trusting in the “Eye in the Sky”? Farmers’ and Auditors’ Perceptions of Drone Use in Environmental Auditing

Worldwide, the agricultural sector is under pressure to demonstrate environmental sustainability. In New Zealand, farm environment plans (FEPs) and their auditing were intended to guide farmers towards sustainable practices by meeting regulations. However, on-farm audits can be time consuming, costly, and stressful for farmers. Meanwhile, the advancement of drone technology has made it possible to incorporate such tools in environmental audits. By means of field observation and in-depth interviews with both farmers and auditors, this research investigated the processes and perceptions of incorporating drones in environmental audits. The aerial views provided additional, high-quality information for the audit. However, flying a drone is subject to weather conditions. Additionally, reductions in audit time were dependent on farm scale, topography, and the auditor’s knowledge of the farm and the farmer. Farmer-auditor relationships are critical for enabling the benefits of drone use within the FEP audit process. Such relationships require a high level of interaction-based trust between farmers and auditors. Further clarity around the use and ownership of drone images could enhance trust, enabling the benefits of drones in audits to be fully utilised, hence furthering the environmental management and compliance processes towards achieving their objectives of better environmental outcomes.

A High-Resolution (20 m) Simulation of Nighttime Low Temperature Inducing Agricultural Crop Damage with the WRF–LES Modeling System

In Korea, sudden cold weather in spring occurs repeatedly every year and causes severe damage to field crops and fruit trees. Detailed forecasting of the daily minimum or suddenly decreasing temperature, closely related to the local topography, has been required in the farmer community. High-resolution temperature models based on empirical formulas or statistical downscaling have fundamental limitations, making it difficult to perform biophysical application and mechanism explanation on small-scale complex terrains. Weather Research and Forecasting–Large Eddy Simulation (WRF–LES) can provide a dynamically and physically scientific tool to be easily applied for farm-scale numerical weather predictions. However, it has been applied mainly for urban areas and in convective boundary layer studies until now. In this study, 20 m resolution WRF–LES simulation of nighttime near-surface temperature and wind was performed for two cold spring weather events that induced significant crop damages in the apple production area and the results were verified with automatic weather station observation data. The study showed that the maximum mean bias of temperature was −1.75 °C and the minimum was −0.68 °C in the spring, while the root mean square error varied between 2.13 and 3.00 °C. The minimum temperature and its duration significantly affected the crop damage, and the WRF–LES could accurately simulate both features. This implies that the application of WRF–LES, with proper nest-domain configuration and harmonized physical options, to the prediction of nighttime frost in rural areas has promising feasibility for orchard- or farm-scale frost prevention and low-temperature management.

Farm-Scale Greenhouse Gas Emissions’ Decision Support Systems

In this paper, 15 farm-scale Green House Gas-based (GHG-based) decision support (DS) tools were evaluated based on a number of criteria (descriptive evaluation), as well as the parameters requested as inputs and the outputs, all of which are considered important for the estimation procedure and the decision support approach. The tools were grouped as emission calculators and tools providing indicators in terms of more than one pillar of sustainability. The results suggest an absence of automatic consultation in decision support in most of the tools. Furthermore, dairy and beef cattle production systems are the most represented in the tools examined. This research confirms a number of important functionalities of modern GHG-based DS tools.