Assessment of coal mine stockpiled soil quality and its impact on vegetation using laboratory-based techniques and reflectance spectroscopy

Thesis (Ph.D. (Agriculture)) -- University of Limpopo, 2018

Surface coal mining requires good and sound rehabilitation practices to re-establish productive land capability and land use after mine-closure. The vast majority of Mpumalanga’s coal deposits are located below high quality and productive arable land. Impacts on soil and land, associated with surface coal mining can reduce the possibility to re-establish the pre-mining land capability and productive potential. Stockpiled soils are excavated from the ground during mining activities, and piled on the surface of the soil for rehabilitation purposes. These soils are often characterized by low Soil Organic Matter (SOM) content, low fertility, and poor physical, chemical and biological properties, limiting their capability for sustainable vegetation growth. The aim of this study was to assess coal-mine stockpile soil quality and its impacts on vegetation using laboratory techniques and Reflectance Spectroscopy. Firstly, the impact of quality of coal-mine stockpile soils on sustainable vegetation growth and productivity was investigated. Soils were collected at three different depths (surface (0-25cm), mid (150-200cm) and deep (300-350cm)), as well as mixed (equal proportion of surface, mid and deep) from two stockpiles (named stockpile 1: aged 10 and stockpile 2:20 years) at the coal mine near Witbank in Mpumalanga Province, South Africa. Soils were amended with different organic and inorganic fertilizer. A 2 x 4 x 5 factorial experiment in a randomized complete block design with four replications was established under greenhouse condition. A grass species (Digiteria eriantha) was planted in pots with unamends and amended soils under the greenhouse condition at ambient temperatures of 26-280C during the day and 16.5-18.50C at night. Mean values of plant height, plant cover, total fresh biomass (roots, stems and leaves) and total dry biomass were found to be higher in the stockpile 1 than in stockpile 2 soils. On average, plants grown on soils with amendments yielded plant height that was 98.28% higher than plants grown on soil with no amendment. On average, height of plants grown on soil amended with poultry manure and lime was 44.65% higher compared to plants planted on soils amended with NPK + lime, compost and poultry manure. On average, mixed soils had better vegetation growth than soil from the individual depths. In total, dry biomass and plant height of plants grown on mixed soils was 33.56% and 22.34% higher than plants grown on surface, mid and deep soils. Mixing soils changes texture, which might affect other physical properties like water availability, infiltration rate and aeration and, to some extent, chemical properties.Secondly, the effect of soil amendments on enzyme activity of coal-mine stockpile soil was investigated. The activity of β-glucosidase, alkaline phosphatase, acid phosphatase and urease was analysed after harvest of grass species (Digiteria eriantha). The results show significantly high activity for β-glucosidase, alkaline phosphatase and urease when soils were amended with poultry manure + lime. Soils with no fertilizer yielded significantly low enzyme activity compared to soil amended with poultry manure+ lime, NPK + lime, sole application of poultry and in some instances compost application. β-glucosidase, urease and acid phosphatase mean values generally tend to decrease with an increase in soil depth. β-glucosidase activity for surface soil was found to be 18.06% higher than that of mid and deep soil. The stockpile depth plays a major role in biochemical activities of the soil; deep soils, in most cases, have decreased microbial biomass and enzyme activity due to oxygen and moisture availability. The results for the effect of organic and inorganic amendment on stockpile soil showed that on average, alkaline phosphatase activity following the application of poultry manure + lime was 17.69% higher than that of lime + inorganic fertilizers (NPK). On average, the acid phosphatase activity following the application of lime + NPK was 56.33% higher than that of poultry manure + lime, compost, soil with no fertilizer as well as sole poultry manure. Urease activity for soil with no fertilizer was found to be 84.70% lower than that of soil amended with poultry + lime. The increase in enzyme activity was attributed to change in soil pH due to application of amendments. A comparison of the two stockpiles indicated that, stockpile 2 (20-year old) had low enzyme activity compared to stockpile 1 (10-year old). The activity of β-glucosidase, acid phosphatase, alkaline phosphatase and urease was found to be 11.03%, 8.04%, 10.03% and 60.23% respectively, higher on stockpile 1, relative to stockpile 2 soils. When soils are stockpiled for a long period of time, microbial biomass is reduced and that affect enzyme activity because microbial biomass is considered as the primary source of enzymes in the soil. Thirdly, the capability to estimate coal-mine stockpile soil properties using Reflectance Spectroscopy was investigated. Soil from coal-mine stockpiles were air dried, crushed, sieved and analysed using laboratory methods. The following soil properties: exchangeable calcium (Ca), sodium (Na), magnesium (Mg), potassium (K), soil pH, organic carbon (OC), phosphorus (P) and clay content were analysed as they are important for vegetation re-establishment during rehabilitation. Spectral reflectance of the soil samples was measured using FieldSpec 3 Portable Analytical Spectral Device (ASD®) spectrometer. Partial Least Square Regression (PLSR) was used to estimate various soil properties, in combination with various spectral transformation techniques such as untransformed reflectance spectra, First Derivative Reflectance (FD) and Log transformed spectra Log (1/R). To assess the performance of various predictive models, R2 (Coefficient of Determination), Root Mean Squares Error of Validation (RMSEV) and Variable Importance in the Projection (VIP) values were computed. The results showed that pH and Ca were accurately estimated (R2=0.79 and 0.69 and RMSEV=0.52 and 0.89cmol/kg respectively) using Log (1/R) reflectance as compared to other soil properties achieving R2 less than 0.5. Ca has strong correlation with pH. Ca expressed in soil solutions is mostly related to pH, which is what was attributed to accurate prediction of both Ca and pH. Soil pH in most cases is directly influenced by calcium carbonate content in the soil. Although the performance of other soil properties was poor, they were highly correlated with pH and Ca except for K. K is soluble and mobile and is therefore subject to leaching in most soils resulting in low K concentrations. Low K concentrations results in higher variability and lower R2 values.Finally, the capability of Partial Least Square Regression and Reflectance Spectroscopy to estimate the effect of coal-mine stockpile soil on foliar nitrogen and phosphorus content was investigated. Grass samples were collected from coal-mine stockpile soils and the adjacent unmined soils at open-cast coal mine around Witbank area in Mpumalanga Province, South Africa. Samples were oven dried and analysed for foliar N and P concentration in the laboratory. Spectral reflectance of the dried grass samples were measured using Analytical Spectral Device (ASD) - FieldSpec 3. Partial Least Square.Regression (PLSR) was used to estimate N and P concentration, in combination with various spectral transformation techniques such as First Derivative Reflectance (FDR) and Log transformed spectra Log (1/R). The results show that stockpile soils appear to impact foliar N and P concentration as evidenced by low N and P concentration in the grass, sampled from stockpile soils compared to grass sampled from unmined soils. This was attributed to soil nutrient status of the study sites, as unmined sites had high soil nutrient content than stockpile soils. Foliar N concentration of grass sampled from stockpile soils and unmined soils can accurately be estimated without spectral transformation. FD yielded highest R2 for N and P estimation in grass sampled from both stockpile soils and unmined soils.Overall, the study shows that stockpiling affect soil quality, enzyme activity and vegetation growth. It further shows that soil amendments can improve soil quality and enzyme activity of coal-mine stockpile soils. Finally, Reflectance Spectroscopy can be used to estimate coal-mine stockpile soil properties, its quality and foliar N and P content as an indicator of vegetation nutrient stress.

National Research Foundation