Dear Colleagues,

Welcome to the new academic year! As many of us returning to our workplace, it’s time to reflect on our achievements during last few months and know more about the extraordinary work of our colleagues. The following are details of some of the key projects undertaken by various research groups in Ontario.

Wishing you all good luck!

Dr. Pradeep Kumar Goel
Ontario Regional Director

Awards/Recognitions

Dr. Prasad Daggupati is the recipient of the 2022 Young Engineer Award from the Northeast Agricultural and Biological Engineering Conference (NABEC) during the annual meeting in Edgewood, Maryland which was held between July 31-August 3, 2022. Dr. Daggupati also received the 2021 young engineer of the year award from the Canadian Society for Bioengineering during the CSBE/SCGAB annual meeting in Charlottetown, PEI which was held between July 24-27, 2022.

Research Achievements

Project # 1: A model for predicting how the frost-free days are increasing 

Researchers at the University of Guelph’s Water Resources Engineering group developed a unique approach to predict the number of frost-free days (FFDs), i.e., winter days when minimum daily temperature is above 0°C. The approach, which is simple yet powerful, estimates number of FFDs by using the rate of increase in observed winter daily minimum temperatures due to climate change.

The data used in this model was collected between 1940 and 2009 from 11 weather stations across Central Canada. The study highlights an overall average temperature increase of about 2°C over 100 years, and an average increase of 12 FFDs in the same time period, for the investigated stations. FFDs each winter were observed to rise exponentially even with linear increases in mean winter daily minimum temperature values. The rate of increase in FFDs varied considerably (0.6 to 36 days per 100 years) among the selected stations, with stations situated at southern latitudes exhibiting relatively greater increases in FFDs, as compared to stations situated in northern areas. The study's main takeaway is that even with a small rise in average daily minimum temperature values in the winter owing to climate change, the number of frost-free days increases significantly, which is alarming. This will have a huge impact on Canada’s agricultural sector as well as winter hydrology.

The approach used to develop this model can be used to estimate the number of days above or below a user-defined threshold temperature (a temperature-dependent variable) in the future, under changing climatic conditions, and thus has huge potential to be used as a decision support tool by various industries (agriculture, energy, urban planning, cities, and regional governments).

Publication Details: Ramesh P. Rudra, Trevor Dickinson, Jaskaran Dhiman, Shaukat Manzoor, Pradeep Goel, Rituraj Shukla. 2022.  A model for predicting how the frost-free days are increasing.  Journal of the ASABE (In Press).

Project #2: Measuring soil organic matter and soil moisture content from digital camera images-comparison of regression and machine learning approaches

Study details:

The overall ecosystem's health is maintained and enhanced through proper soil management. The adequate characterisation of soil parameters, such as soil organic matter (SOM) and soil moisture content (SMC), is necessary for effective management of the soil. In the study, SOM and SMC were predicted using digital camera images. Compared to conventional approaches, image-based soil characterisation has demonstrated significant potential. SMC and SOM are known to affect soil colour. Higher light absorption is the cause of the darker appearance associated with higher moisture content. However, with time, increased anaerobic conditions and the state of iron oxides in the soil can have an impact on soil colour (Jackson 2008).

Developing a cell phone app that can be used by anyone with a smartphone to estimate key critical soil parameters such as SOM and SMC offer the potential for more affordable, quick, and easy-to-use soil analysis, helpful for supporting management decisions.

In the study, the processed soil samples were evenly placed in Petri dishes (∼8mm thickness) and the surface of the samples were captured with a 12.1-megapixel digital camera (Canon PowerShot SX270 HS) mounted on a tripod (27 cm) with the lens facing downward toward the sample. In this study, 22 supervised regression and machine learning algorithms were calibrated and validated in order to assess how well soil photographs taken with a digital camera could predict SOM and SMC. These models created correlations between numerous color- and texture-related variables extracted from photos, SOM and SMC (measured in the lab).

Color parameters demonstrated high correlation with both SOM and SMC. Overall, SMC was predicted more accurately than SOM, indicating that SMC has a significant influence on the soil's hue. For the validation dataset using 6 predictor factors, the results showed a reasonable agreement between the image parameters and the laboratory measured SOM and SMC. For this investigation, the non-linear interactions between SOM, SMC, and image properties were best captured and explained by GPRs and tree models (Cubist, RF, and Boosted Trees). The soil colour is also influenced by temperature, climate, and mineral content. The image parameter based, and corresponding laboratory results revealed that that SOM and SMC can be predicted with more than 75% accuracy. The study also revealed that model built using the key image parameters is more rapid and can produce result as good as built using all the parameters. This methodology's advantage over the conventional approach would be quick assessment of soil parameters at a significantly lower cost and environmentally safe. Taken together, digital image-based soil characterization offers a chance to be applied to proximal soil sensing.

Study highlights:

• Soil organic matter (SOM) and moisture content (SMC) were predicted using digital camera images.
• Color and textural parameters had a high correlation with SOM and SMC.
• A satisfactory agreement found between the image parameters and the laboratory measured SOM and SMC.
• GPRs and tree models best captured the non-linear relationships between SOM, SMC, and image parameters.

Publication Details: Perry Taneja, Hiteshkumar Bhogilal Vasava, Solmaz Fathololoumi, Prasad Daggupati, and Asim Biswas. 2022. Predicting soil organic matter and soil moisture content from digital camera images: comparison of regression and machine learning approaches. CJSS: https://doi.org/10.1139/cjss-2021-0133