Quantifying soil organic matter in urban soils using visible and near-infrared spectroscopy in Zagreb city

master's thesis

• Title (croatian): Kvantifikacija organske tvari u urbanim tlima na području grda Zagreba primjenom vidljive i blisko-infracrvene spektroskopije
• Author: Sabina Poudel
• Mentor: Monika Zovko (mentor)
• Mentor: Martin Kulhanek https://orcid.org/0000-0003-2755-7215 (mentor)
• Committee member: Monika Zovko (predsjednik povjerenstva)
• Committee member: Martin Kulhanek https://orcid.org/0000-0003-2755-7215 (član povjerenstva)
• Committee member: Ivana Šestak (član povjerenstva)
• Committee member: Aleksandra Bensa (član povjerenstva)
• Granter: University of Zagreb; Faculty of Agriculture; (Soil Amelioration); Zagreb
• Defense date and country: 2024, Croatia
• Scientific / art field, discipline and subdiscipline: BIOTECHNICAL SCIENCES; Agronomy

Abstract

Soil organic matter (SOM) is a fundamental constituent of soil, and its accurate quantification
is essential for understanding anthropogenic changes and guiding effective urban management.
Visible and near-infrared (Vis-NIR) spectroscopy has been utilized to measure SOM content in
diverse land use system offering a time and cost-effective method. However, limited research
has been conducted to explore the relationships between SOM and spectral data in urban soils.
This study aims to evaluate the capability of Vis-NIR spectroscopy to assess SOM in the urban
environment of Zagreb, addressing the specific challenges posed by urban areas and developing
a spectral model dedicated to accurately predicting SOM changes in these soils. A total of 150
soil samples were collected from various land-use systems within the city of Zagreb and
analysed for SOM content using the Walkley-Black method. The reflectance spectra were
measured using a Vis-NIR spectroradiometer (Spectral Evolution SR2500, USA) instrument in
the range from 350 to 2500 nm in the laboratory. Prediction models for SOM content were
developed using partial least square regressions (PLSR) combined with reflectance spectrum
pre-processing methods, including band clipping, Savitzky-Golay smoothing, and Standard
Normal Variate (SNV) transformation. Key wavelengths for predicting SOM were identified
for visible region 350 to 700 nm and for near-infrared region 700 to 2500 nm. The models
achieved an r-squared (R²) of 0.96, root mean square error (RMSE) of 0.35, ratio of
performance to deviation (RPD) of 5.40 demonstrating Vis-NIR spectroscopy's reliability in
predicting SOM in urban soils, supporting better urban planning and environmental monitoring.

Abstract (croatian)

Organska tvar tla (SOM) temeljni je sastojak tla, a njezina točna kvantifikacija ključna je za
razumijevanje antropogenih promjena i učinkovitog upravljanja urbanim zemljištem. Vidljiva i
bliska infracrvena (Vis-NIR) spektroskopija korištena je za određivanje sadržaja SOM u
različitim sustavima korištenja urbanog zemljišta, nudeći vremenski i troškovno učinkovitu
metodu. Dosada su provedena ograničena istraživanja kako bi se istražili odnosi između SOM
i spektralnih podataka u urbanim tlima. Ova studija ima za cilj procijeniti sposobnost Vis-NIR
spektroskopije za procjenu SOM u urbanom okruženju Zagreba, rješavajući specifične izazove
koje predstavljaju urbana područja i razvijajući spektralni model za predikciju SOM u tim tlima.
Ukupno je prikupljeno 150 uzoraka tla iz različitih sustava korištenja zemljišta na području
grada Zagreba i analizirano na sadržaj SOM-a Walkley-Black metodom. Spektri refleksije
mjereni su instrumentom Vis-NIR spektroradiometrom (Spectral Evolution SR2500, SAD) u
rasponu od 350 do 2500 nm u laboratoriju. Modeli predviđanja za sadržaj SOM-a razvijeni su
korištenjem djelomičnih regresija najmanjih kvadrata (PLSR) u kombinaciji s metodama
prethodne obrade spektra refleksije, uključujući isječak spektralnih vrpci, Savitzky-Golayevo
izglađivanje i transformaciju standardne normalne varijacije (SNV). Ključne valne duljine za
predviđanje SOM identificirane su za vidljivo područje od 350 do 700 nm i za blisko
infracrveno područje od 700 do 2500 nm. Modeli su postigli R² (R-kvadrat) od 0,96, srednju
kvadratnu pogrešku (RMSE) od 0,35, i omjer performansi i adstupanja (RPD) od 5,40,
pokazujući pouzdanost Vis-NIR spektroskopije u predviđanju SOM-a u urbanim tlima,
podržavajući bolje urbano planiranje i praćenje okoliša.

• Keywords: Soil quality; Standard Normal Variate (SNV); Savitzky-Golay (SG)
• Keywords (croatian): kvaliteta tla; standardna normalna varijanta (SNV); Savitzky-Golay (SG)
• Language: english
• URN:NBN: urn:nbn:hr:204:015514
• Study programme: Title: Sustainability in Agriculture, Food Production and Food Technology in the Danube Region; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka agronomije (magistar/magistra inženjer/inženjerka agronomije)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Access restricted to students and staff of home institution
• Created on: 2024-09-05 08:54:28