Search

The production of wild blueberries (Vaccinium angustifolium) contributes 112.2 million dollars yearly to Canada's revenue, which can be further increased by reducing harvest losses. A precise prediction of blueberry harvest losses is necessary to mitigate such losses. The performance of three machine learning (ML) algorithms was assessed to predict the wild blueberry harvest losses on the ground. The data from four commercial fields in Atlantic Canada (including Tracadie, Frank Webb, Small Scott, and Cooper fields) were utilized to achieve the goal. Wild blueberry losses (fruit loss on ground, leaf losses, blower losses) and yield were measured manually from randomly selected plots during mechanical harvesting. The plant height of wild blueberry, field slope, and fruit zone readings were collected from each of the plots. For the purpose of predicting ground loss as a function of fruit zone, plant height, fruit production, slope, leaf loss, and blower damage, three ML models i.e., support vector regression (SVR), linear regression (LR), and random forest (RF)—were used. Statistical parameters i.e., mean absolute error (MAE), root mean square error (RMSE), and coefficient of determination (R2), were used to assess the prediction accuracy of the models. The results of the correlation matrices showed that the blueberry yield and losses (leaf loss, blower loss) had medium to strong correlations accessed based on the correlation coefficient (r) range 0.37–0.79. The LR model showed the foremost predictions of ground loss as compared to all the other models analyzed. Tracadie, Frank Webb, Small Scott, and Cooper had R2 values of 0.87, 0.91, 0.91, and 0.73, respectively. Support vector regression performed comparatively better at all the fields i.e., R2 = 0.93 (Frank Webb field), R2 = 0.88 (Tracadie), and R2 = 0.79 (Cooper) except Small Scott field with R2 = 0.07. When comparing the actual and anticipated ground loss, the SVR performed best (R2 = 0.79–0.93) as compared to the other two algorithms i.e., LR (R2 = 0.73 to 0.92), and RF (R2 = 0.53 to 0.89) for the three fields. The outcomes revealed that these ML algorithms can be useful in predicting ground losses during wild blueberry harvesting in the selected fields. [ABSTRACT FROM AUTHOR]

Modelling the distribution and abundance of species at risk is extremely important for their conservation and management. We used ecological niche models (ENMs) to predict the occurrence of western rattlesnakes (Crotalus oreganus) in British Columbia (BC), Canada. We applied this to existing population estimates to support a threshold of occurrence for management and conservation. We also identified predictors influencing rattlesnake distribution and abundance in this region. Using a Geographic Information Systems platform, we incorporated ENMs, capture–mark–recapture (CMR) and radio-telemetry results, province-wide observations, Landsat imagery and provincial databases for agricultural land use to produce quantitative, spatially explicit, population estimates across BC. Using available western rattlesnake habitat estimated at 183.9 km2 and averaging estimates calculated from densities in three study populations, we generated a mean adult population size of 9722 (±SD 3009; 0.8 relative index of occurrence [RIO] threshold). Only a small area (21.6 km2) of suitable land cover was located within protected areas, potentially protecting an estimated 1144 (±354) adults. Most suitable land cover was within 500 m of roads (170.6 km2), representing potential habitat being used by an estimated 9017 (±2791) adults. At the threshold RIO value chosen (0.8), only a very small area of farmland provided suitable land cover. Our results highlight the possibility of high mortality rates for western rattlesnakes near roads and the fact that protected areas do not provide sufficient coverage to conserve the population. Given that this species has relatively low mobility and high site fidelity to home ranges, our population estimate for BC provides a useful reference for the northern part of the species' range. It also fulfills a need to estimate population size within political jurisdictions where conservation management decisions are made, as well as presenting a method that can be applied to other parts of the range, including the southern United States. Our study provides an important benchmark for future monitoring of western rattlesnakes in BC using a repeatable and transparent approach. Similar applications can be extrapolated and applied for other threatened species to identify and quantify population distributions and threats, further supporting conservation prioritization tools to be used to maximize the effectiveness of conservation strategies under financial constraints. [ABSTRACT FROM AUTHOR]

This research aimed at exploiting the joint use of machine learning and polarimetry for improving the retrieval of surface soil moisture content (SMC) from synthetic aperture radar (SAR) acquisitions at C-band. The study was conducted on two agricultural areas in Canada, for which a series of RADARSAT-2 (RS2) images were available along with direct measurements of SMC from in situ stations. The analysis confirmed the sensitivity of RS2 backscattering (σ°) to SMC. The comparison of SMC with the compact polarimetry (CP) parameters, computed from the RS2 acquisitions by the CP data simulator, pointed out that some CP parameters had a sensitivity to SMC equal or better than σ°, with correlation coefficients up to R ≃ 0.4. Based on these results, the potential of machine learning (ML) for SMC retrieval was exploited by implementing and testing on the available data an artificial neural network (ANN) algorithm. The algorithm was implemented using several combinations of σ° and CP parameters. Validation results of the algorithm with in situ observations confirmed the promising capabilities of the ML techniques for SMC monitoring. Furthermore, results pointed out the potential of CP in improving the SMC retrieval accuracy, especially when used in combination with linearly polarized σ°. Depending on the considered input combination, the ANN algorithm was able to estimate SMC with Root Mean Square Error (RMSE) between 3% and 7% of SMC and R between 0.7 and 0.9. [ABSTRACT FROM AUTHOR]