Remote Sensing of Grassland Variables Across Seasons and Using Multiple Spectral Devices

Pearce, Craig and Gerard, France and Blackburn, George and Harris, Paul and Smart, Simon (2023) Remote Sensing of Grassland Variables Across Seasons and Using Multiple Spectral Devices. PhD thesis, Lancaster University.

[thumbnail of 2023pearcephd]
Text (2023pearcephd)
2023pearcephd.pdf - Published Version

Download (16MB)

Abstract

The regeneration and conservation of semi-natural grasslands is considered important to land managers such as Natural England, especially grasslands protected by legislation such as UK Biodiversity Action Plan (BAP) priority habitats or Sites of Special Scientific Interest (SSSI). Monitoring the condition of these grasslands is necessary, but conventional methods of measuring grassland condition are time consuming and limited in their spatial coverage. This thesis tested the hypothesis that remote sensing (RS) techniques can provide a cost- and time-effective solution to grassland condition monitoring. This thesis used partial least squares regression (PLSR) to explore the relationship between grassland spectral reflectance and the mass or % cover of a range of condition-related grassland variables plus a metric (an average and equally weighted measure of whether the CSM criteria were sufficiently met referred to as CSMcondition) representing condition as defined in the UK by the Common Standards Monitoring. The relationship between grassland variables and CSM-condition was also assessed. Each study differed with the grasslands targeted, the seasons when data were collected and the devices deployed. The first study was conducted on a range of different grassland types, the second study was conducted on chalk grasslands of differing levels of improvement across three seasons (spring, summer and autumn) and the third study was conducted on these same chalk grasslands but using data from three different spectral devices collected during the summer. All three studies were conducted at patch level (1m2) with the third study including the extrapolated predictions from trained statistical models to field level (200x1m) using spectral data from a CROPSCAN MSR 16R hand-held device. All three studies used spectral data from a CROPSCAN MSR 16R hand-held device and the third study included the analysis of spectral data from a Spectral Vista Corporation (SVC) HR1024i hand-held device and a Rikola camera mounted on an uncrewed aerial vehicle (UAV). The results suggest that some of the condition-related variables considered in this thesis are predicted with reasonable accuracy and precision at patch level, but producing reliable results requires a sufficient quantity of data to train the statistical models (at least 30 quadrats of samples in the context of this thesis) especially if the results are to be extrapolated to field level as additional data are required for the external validation of the results. When analysing data collected at patch level during the summer; the mass of bryophytes, dead material and graminoids plus the % cover of forbs can be predicted to a moderate level of accuracy when analysing data from all seven grasslands. When analysing data from all Parsonage Down NNR grasslands; the mass of bryophytes, the % cover of live material, % cover-based live:dead ratio and CSM-condition could be predicted to a high level of accuracy. Moisture content plus the % cover of dead material, forbs and gram:forb ratio were all predicted to a moderate level of accuracy as well as CSM-condition predicted by grassland variable values. When using data from all Ingleborough NNR grasslands; the % cover of forbs and biomass plus the mass of bryophytes, dead material and live material could be predicted to a moderate level of accuracy. When using patch level data collected across three seasons; the % cover of dead material, live material and live:dead ratio plus the mass of graminoids could be predicted when using three seasons of data collected on one grassland, or for all three Parsonage grasslands, to at least a moderate level of accuracy although some models trained with % cover data had a high accuracy. Forbs (mass and % cover) plus the mass of gram:forb ratio, live material and live:dead ratio could be predicted to at least a moderate level of accuracy for some grasslands. When using data from all grasslands collected in one season to predict grassland variables; the mass of a range of grassland variables could be predicted to a moderate level of accuracy for the spring and autumn months but not when using % cover data. When the use of data from three different spectral devices were compared to see which produced the most accurate models; using CROPSCAN and SVC data produced similar results, with slightly stronger results from the CROPSCAN, but using data from the Rikola camera produced weaker results. When the results of trained PLSR models were extrapolated to field level, the projected predicted grassland variable values from models trained with CROPSCAN MSR 16R data looked promising but the results have not been externally validated using a separate data set. Variable importance in projection (VIP) was used to establish which spectral bands are most important for predicting each grassland variable plus CSM-condition and which grassland variables are most important in predicting CSM-condition. It was generally found that the upper parts of the spectral range of each device (NIR and SWIR) were the most crucial for predicting grassland variables, with the red edge (647nm) and particular visible bands (470nm) also having some importance. When grassland variables were used to predict CSM-condition, which variables were most important depended on whether the grassland variable was mass-based or % cover-based. When using mass data; graminoid:forb ratio mass and live:dead ratio masswere consistently important across grasslands and seasons with biomass, graminoid:bryophyte mass and moisture content having importance for particular grasslands and seasons. When using % cover data; forbs cover, graminoids cover and live:dead ratio cover were consistently important across grasslands and seasons with dead material cover and live material cover having importance for particular grasslands and seasons. This thesis also explored which grassland variables could be predicted most consistently by calculating coefficient of variance (CV) on data collected across grasslands, seasons and/or using different spectral devices. Overall, these results suggest that none of the grassland variables considered in this thesis can be consistently predicted strongly across all the different grasslands or seasons considered in this thesis. When using % cover variable data; forbs cover and live:dead ratio cover produced relatively consistent results across grasslands, seasons and when using data from different spectral devices while bryophytes cover, graminoids cover and gram:forb ratio cover were consistent under some specific circumstances. When using mass data; moisture content stands out as relatively consistent compared to other variables across grasslands, seasons and when using different spectral devices. When using CROPSCAN MSR 16R spectral data as predictors, live material mass and live:dead ratio mass plus biomass produced relatively consistent results. Dead material mass produced relatively consistent results when using different devices as predictors, but not when using data collected over different seasons

Item Type:
Thesis (PhD)
Subjects:
?? remote sensinggrassland conditionplsruav ??
ID Code:
185617
Deposited By:
Deposited On:
03 Feb 2023 11:00
Refereed?:
No
Published?:
Published
Last Modified:
20 Dec 2023 06:39