Junxiang Peng
Research Scientist
(+47) 403 40 352
junxiang.peng@nibio.no
Place
Særheim
Visiting address
Postvegen 213, NO-4353 Klepp stasjon
Abstract
Potato field management in Europe is already optimized for high production and tuber quality; however, numerous environmental challenges remain if the industry is to achieve “green economy” targets, such as less resources utilized, and less nitrate leached to the environment. Strategic co-scheduling irrigation and nitrogen (N) fertilization might increase resource use efficiency while minimizing reactive losses such as nitrate leaching. This study aimed to quantify the combined effect of irrigation and N fertilization on potato production, growth, and resource use efficiencies. A field experiment was conducted from 2017 to 2019 on a coarse sandy soil in Denmark, with a drought event occurring in 2018. Full (Ifull, maximized), deficit (Idef, 70–80 % of Ifull) and low irrigation treatments (Ilow, minimized amount to keep crop survival), each under full (Nfull, maximized) and variable (Nvar, variable amount according to the crops’ needs) N fertilization were applied. The analyses results show that Ilow limited potato growth under a drought-heat event; otherwise, potato growth was comparable between Ifull and Idef treatments, with 31–32 % higher irrigation efficiency (IE) under Idef than under Ifull. Nitrate leaching was variable and not significantly different among the treatments, being in general 9–13 % lower under Idef in absolute terms than under Ifull. Unexpectedly, outcomes from Nvar were statistically lower compared to those from Nfull. Radiation use efficiencies (RUEs) from Ilow and Nvar were significantly lower than from Ifull and Idef (14–19 %), and from Nfull (9–11 %). N use efficiencies (NUE) were comparable between N fertilization treatments but significantly different among different irrigation treatments. Overall, this study confirms that Idef is the best irrigation strategy. Future efforts should focus on developing improved approaches for detecting in-season crop N status and further quantifying N requirements, as well as promoting the co-scheduled management of irrigation and N fertilization. Remote sensing approaches have great potential to assist with this.
Authors
A. Terskaia R. Oucheikh Junxiang Peng J. Oliveira D. Parsons J. Morel A.M. Abdi Z. Cai S. Thapa N. Boke-Olen P. Olsson H.G. Smith J. Ekroos L. EklundhAbstract
Grassland management practices are critical both for maintaining resilient food production and for preserving biodiversity, yet spatially explicit information on them is often limited. Recent advances in freely available, remotely sensed land surface phenology and productivity datasets raise the question of whether phenological metrics provide additional, biologically interpretable information for characterising and mapping grassland management beyond conventional spectral features. In this study, we evaluated the information content and added value of freely available, ready-to-use Sentinel-2–derived High-Resolution Vegetation Phenology and Productivity (HR-VPP) data for distinguishing broad grassland management types (conventional leys, organic leys, and semi-natural grasslands) across multiple spatial scales. We analyzed 1455 grasslands across the intensively farmed Skåne province (southern Sweden) over 2018–2022 at provincial and local scales. We further tested whether phenological variables improve grassland management classification when combined with multi-temporal Sentinel-2 spectral data using state-of-the-art machine learning (ML) methods, and applied a conformal prediction framework to manage classification uncertainty. Relationships between HR-VPP metrics and field-measured biomass were additionally assessed using independent field data from northern Sweden. Field-level analyses showed weak (R2 = 0.28) to moderate (R2 = 0.47) relationships between HR-VPP productivity metrics and measured biomass. HR-VPP metrics captured statistically significant but subtle differences among management types, with strong overlap between categories, particularly at the provincial scale. Differences were more pronounced at local scales. Consequently, HR-VPP variables alone had limited predictive power and ML model accuracy was low (0.53). However, combining HR-VPP with Sentinel-2 spectral time-series data substantially improved classification performance, achieving an accuracy of 0.70. Feature importance analysis indicated that spectral indices dominated classification, with limited direct contribution from phenological parameters. These findings highlight both the potential and current limitations of continental-scale phenology and productivity products for grassland management monitoring and point to the need for improved models and complementary data sources to better capture management practices.
Abstract
Nitrogen (N) management is one of the main factors enhancing potato productivity and promoting sustainable agricultural practices. The Nitrogen Nutrition Index (NNI, obtained as the ratio of actual plant N, to the critical plant N concentration) is widely applied to assess the N status of various crops. Traditionally, NNI is calculated using field data, but remote sensing (RS) technologies can offer more rapidly and timely assessment of the spatiotemporal (within field) variability of this index. This study employs multispectral data acquired via Unmanned Aerial Vehicle (UAV) and machine learning (ML) models to estimate potato NNI. A Bayesian hierarchical partially pooled method was fitted to a three-year f ield experiment in Denmark and extensive ground-based potato datasets to model the critical nitrogen dilution curve (CNDC) and calculate the NNI. Multispectral UAV data were processed to extract four spectral bands and calculate several vegetation indices, which were used as predictors to train and test six ML models: Linear regression, support vector machines, gaussian process regression, stepwise linear regression, ensemble trees and neural networks. Among the compared models, gaussian process regression outperformed, showing R 2 equal to 0.83 and a RMSE of 0.10 and providing accurate NNI predictions, comparable to ground-based Bayesian estimates. The variability of the NNI was analyzed over the seasons using 28 NNI maps derived from UAV surveys at spatial resolution of 0.04–0.09 m/pixel, capturing spatial variations in crop N status over time. The proposed framework, designed for NNI prediction at the intra-field scale, has the potential to be adapted to different environments and crops. The framework can support practical decisions for precision N management, reducing the environmental impact of potato cultivations and enhancing sustainability.
