Irrigation and fertilization, particularly with nutrients such as potassium, phosphorus, calcium, and magnesium, have significant effects on potato tubers. In this study, the effects of different irrigation levels and fertilization practices on the yield and quality parameters of stored potatoes were investigated. The research was conducted as part of a 2-year trial (2019–2020) at the Application Field of the Yenişehir İbrahim Orhan Vocational School, Uludağ University, Bursa. Potato storage was carried out in a storage facility located at the Vocational School Research Field. Four different irrigation levels (I100, I75, I50, and I25) and three different fertilization levels (F1, F2, and F3) were examined for their effects on yield and quality parameters of stored potato tubers. During the 2-, 4-, and 6-month storage periods at temperatures of 8–10 °C and relative humidity of 95
The increasing global population and food security concerns are rapidly raising the demand for sustainable approaches to increasing agricultural productivity. Triacontanol (TRIA) is a non-toxic primary alcohol that can be naturally obtained from plant waste and has regulatory effects on growth, yield, and physiological processes in plants. Investigating the contribution of such natural growth regulators, which are low-cost and accessible to farmers, to yield increase in potato farming is of great importance for sustainable agriculture. This study was conducted in a total of 270 pots under greenhouse conditions using a randomized block design with 5 different TRIA doses, 6 different application methods (control, IAA, SA, GA3, ABA, and mix) and 3 different application approaches (foliar, soil, and soil + foliar). The study aimed to determine the effects of TRIA on the Agria potato variety (Solanum tuberosum L.) and to identify the most effective combination of application methods for yield enhancement. According to the results, foliar application of TRIA at a dose of 10 µM yielded the most effective outcomes compared to other methods and doses. Under greenhouse conditions, foliar-applied 10 µM led to a 122.97
Potato (Solanum tuberosum) is the fifth most important food crop globally, with an average per capita consumption of 32.9 kg. In Latin America, it ranks fourth among staple crops, after maize, rice, and wheat. In Bolivia, potato is a key dietary component, sustaining the population whose per capita consumption of 100 kg ranks sixth globally and first in South America. This study analyzes the growth rate, trends, and forecasts of potato cultivation in Bolivia, focusing on national and regional patterns in cultivated area, yield, and production to identify underlying drivers and provide insights into the factors influencing potato productivity. The analysis covers the period from 1961 to 2022 for the national-level data and from 1984 to 2022 for the regional-level data. Results revealed that the total cultivated area of potato in Bolivia increased at an average annual growth rate of 1.47
Early plant disease detection is essential for sustainable crop management and yield optimization. However, environmental variability, noise in the images and blur often reduce the accuracy of conventional deep learning models in the field conditions. To address these limitations, we propose a novel hybrid texture-based deep learning framework. In this work we integrate local ternary pattern (LTP), local phase quantization (LPQ) and convolutional neural networks (CNNs) through transfer learning. The LTP captures local edge and texture. The LPQ captures the phase of information. The pretrained CNN backbone extracts high-level semantic features. The combination of these features can enhance classification accuracy in real-world environments. The experiment is conducted on a dataset of seven potato leaf disease classes. These datasets are collected in an uncontrolled environment. The proposed framework achieved a good classification average accuracy of 94.65
A 2-year field study was conducted in Chikkaballapur, Karnataka, India, during the consecutive rabi seasons of 2022–2023 and 2023–2024 to evaluate the physiological growth, yield dynamics, and economic viability of potatoes (Solanum tuberosum L.) under integrated nutrient regimes that combine foliar applications of nano-fertilizer with titrated soil macronutrients. The experiment had 14 treatments replicated thrice and set up in a randomised complete block design to evaluate various soil nitrogen (N) and potassium (K) levels paired with sequential foliar sprays of nano-N and nano-K. The results demonstrated that the use of nano-fertilizers in place of some conventional soil inputs significantly enhanced morpho-physiological traits, including tuber distribution metrics, plant height, and SPAD chlorophyll index values. The treatment that received three foliar sprays of nano-N + K at 30, 45, and 60 days after planting (DAP) (T14) in addition to 75
Potato (Solanum tuberosum L.) is a strategic crop for food security, income generation, and livelihood diversification in the Ethiopian highlands. Despite its historical importance as a major seed potato production area, Degem District has experienced a sustained decline in productivity over the last two decades. This study investigated the drivers of potato production change in Degem District, Central Ethiopia, using a Structural Equation Modeling (SEM) framework. Primary data were collected from 550 potato-producing households through structured surveys, complemented by focus group discussions, key informant interviews, and secondary datasets covering the period 2005–2025. The SEM approach was employed to examine the complex interactions among climatic, biophysical, institutional, market, and socioeconomic factors affecting potato productivity. The findings revealed that disease pressure, climate variability, and seed system constraints were the most influential exogenous latent constructs contributing to productivity decline. Average potato yield declined to 10–15 t ha−1, representing an estimated productivity reduction of approximately 28.6
Potato, the fourth most important staple crop worldwide, is highly susceptible to weed interference, which substantially reduces both yield and quality. To enable precise and intelligent weed management, a Down-Scale Attention Potato-Weed NetWork (DSAP-WeedNet) model was developed based on YOLOv8n. An adaptive downsampling module replaced the conventional downsampling layers to enhance the preservation of fine-grained feature details. In addition, a multi-path coordinate attention module was incorporated between the neck and the detection head, effectively integrating global semantic information with local spatial cues, thereby improving feature discrimination and localization accuracy. Experiments were conducted on a dataset comprising 2055 images of potato plants and field weeds. DSAP-WeedNet achieved an mAP50 of 97.4 F_1 score of 97.65
Potato (Solanum tuberosum L.) is an important global crop, and optimizing nutrient management in aeroponic systems is crucial for efficient mini-tuber production. This study evaluated the effects of Hoagland nutrient solution concentration (HSC) (25
The adoption of climate-resilient potato varieties (CRPVs) in Kenya remains limited despite sustained investments in breeding for tolerance to drought, heat, and emerging pests and diseases. Existing studies frequently explain this gap through isolated constraints such as low farmer awareness, limited seed access, or poor varietal performance. This review instead conceptualises low adoption as a consequence of systemic misalignment across breeding priorities, seed systems, market preferences, and governance structures within potato innovation system. Using a system-based perspective, the study synthesises interdisciplinary literature and policy evidence published between 2005 and 2025 through a critical narrative review framework to examine interaction, feedback mechanisms, and institutional dynamics influencing varietal uptake. The findings reveal that although breeding programmes are increasingly climate-responsive, they remain insufficiently aligned with farmers’ multi-stress production realities, livelihood priorities, and evolving market requirements. Adoption is further constrained by fragmented and under-resourced seed systems that limit access to CRPVs, while market structures frequently favour established cultivars possessing preferred processing, trading, and consumer traits. These constraints are reinforced by institutional fragmentation, weak coordination among value-chain actors, and limited integration of market intelligence into breeding processes, creating reinforcing feedback loops that perpetuate low sustained adoption. The review concludes that improving varietal uptake requires moving beyond linear technology dissemination models towards more integrated and adaptive innovation systems that strengthen alignment among breeding programmes, seed systems, markets, and governance processes. It highlights the importance of participatory and demand-led breeding, pluralistic seed system development, stronger market integration, and coordinated multi-level governance for enhancing adoption and impact of CRPVs.
Potato prices are driven by production cycles, retail market circulation, regional co-movement, and recurrent supply–demand fluctuations, making accurate forecasting challenging under price-only conditions. This paper proposes MarketDNA, a discrete market-state foundation model for daily potato price forecasting. Instead of directly regressing future prices from continuous historical observations, MarketDNA reformulates agricultural price forecasting as a latent market-language modeling problem. It first learns a discrete vocabulary of reusable market-state tokens from local price fragments, allowing noisy price trajectories to be represented as compact price-regime sequences. A cross-centre state transformer is then developed to capture temporal state transitions and market co-movement across retail centres. MarketDNA is pretrained through market-language objectives, including market-state reconstruction, masked state modeling, state transition prediction, and cross-centre state alignment, without relying on weather, logistics, inventory, production, storage, policy, or other external variables. Experiments show that MarketDNA achieves competitive short-term forecasting performance and superior multi-step forecasting accuracy compared with representative baselines. Ablation analysis confirms the effectiveness of the discrete tokenizer, market-language pretraining, and cross-centre state transformer. The results suggest that modeling agricultural price dynamics as a discrete market-state language provides an effective and interpretable strategy for masked multi-centre price forecasting.
Lightweight potato disease detectors are promising for field deployment, but their compact backbones often lose the fine spatial evidence required to distinguish early blight, late blight, and healthy leaves. This paper presents NanoTeacher-YOLO26, a foundation-to-edge self-supervised distillation framework that transfers teacher-side visual priors into a deployable YOLO26n student without increasing inference cost. A stronger YOLO26s teacher is used only during training to provide structured disease-aware supervision, including confidence-weighted feature geometry, disease-evidence attention, and boundary-gradient priors. These signals are integrated into a Disease-Aware Distillation Loss, enabling the nano student to inherit the teacher’s localization behavior and spatial response patterns while preserving its original model footprint. Experiments are conducted on a YOLO-format potato leaf disease benchmark derived from the public PlantVillage potato subset. Results show that NanoTeacher-YOLO26 reaches the saturated mAP regime of the benchmark and improves over the same-footprint YOLO26n baseline in precision and recall. Detection overlays, attention maps, teacher-student alignment, confusion analysis, and training curves further indicate that the distilled student learns disease-relevant spatial evidence rather than relying only on class-level shortcuts. The study provides a reproducible route for converting teacher-side perception into edge-ready agricultural detectors.
Black dot has become an increasingly major issue in potato production, with recent studies highlighting its importance. The fungal pathogen responsible for this disease, Colletotrichum coccodes, is widely distributed across major potato-growing regions and in fields where alternative hosts are cultivated. The pathogen has a broad host range, including many economically important crops, several of which rank amongst the most widely produced and high-value commodities globally. Direct yield losses associated with black dot disease in potato production have been documented, with yield reductions ranging from 19 to 48
Potato soft rot, induced by Pectobacterium carotovorum and related species, precipitates severe worldwide economic consequences, frequently causing crop losses of up to 60
Climate change is expected to intensify heat and water stress in semi-arid regions, increasing the need for adaptive, data-driven tools to support sustainable crop production. This study quantified the future blue and green components of the water footprint (WF) of irrigated potato farming in Niğde Province (Türkiye) by integrating field-based calibration and validation of the DSSAT SUBSTOR-Potato model, driven by a bias-corrected multi-model ensemble of 24 General Circulation Models from the CMIP6 archive. Model evaluation showed variable performance across traits, with nRMSE values ranging from 9.9 to 69.8
To enable rapid and non-destructive estimation of potato plant nitrogen content (PNC), this study used hyperspectral data to derive three categories of spectral features, namely Empirical Vegetation Indices (VIs), Spectral Edge Parameters (EPs), and Three-Dimensional Optimal Spectral Indices (3D-OSI). Their correlations with measured PNC were calculated to identify candidate spectral variables sensitive to potato nitrogen status. To avoid relying solely on statistical significance, minimum redundancy maximum relevance (mRMR) analysis was additionally used as a supplementary redundancy check to support feature selection and interpretation. Based on these candidate variables, PNC estimation models were established using Random Forest (RF), Back Propagation Neural Network (BPNN), and Partial Least Squares Regression (PLSR). On the basis of the correlation analysis, the performance of different feature types and their combinations was further compared. The results showed that 3D-OSI had the strongest association with PNC, with the maximum absolute correlation coefficient reaching 0.749, followed by EPs (0.686), whereas the correlations of VIs were relatively weaker (0.510). When each feature category was used separately, 3D-OSI produced the best prediction results, and the RF model based on this feature set achieved an R2 of 0.749. Model performance improved further when different feature types were integrated. Among all input combinations, the RF model using VIs + EPs + 3D-OSI gave the best results, with an R2 of 0.800. Compared with the model developed using VIs alone, the fused-feature-set model increased R2 by 31.58
Potato plays a critical role in achieving the United Nation’s Sustainable Developmental Goal 2 (SDG2) of zero hunger. West Bengal, as the second-largest potato-producing state in India, has a significant role in food security. This micro-level study addressed the need to plug loopholes in the potato supply chain at the farmer level. This research was conducted in two randomly selected villages—Morhal and Mukundapur—of Hooghly district, West Bengal, during 2024–2025, analyzing data from 204 farmers. The statistical methods, including descriptive statistics, association measures, and cluster analysis, were used to examine socio-demographic factors and potato cultivation economics. The results indicated that socio-demographic features like family size, number of adults in the family, land holding, number of crops taken, and cropping intensity impact farm income. Farmers with larger land sizes had comparatively lower farm income per unit of land. We examined that potato contributed 68
One of the fundamental government policies in Cameroon has been the promotion of second-generation agriculture through household food security, self-sufficiency and improved living conditions for rural population. Despite these policies, the level of efficiency in crop production remains below expectations. It is on this prelude that this paper investigated the technical efficiency level and efficiency gap between Irish potato (Solanum tuberosum L.)-producing households in the West and in the North West Regions of Cameroon. To achieve these objectives, the paper employed the non-parametric two-stage Data Envelopment Analysis (DEA) approach, the fractional logit model to ascertain the determinants of technical efficiency, and the Oaxaca-Ransom decomposition framework to measure the efficiency gap between households. A sample of 604 households was selected using the proportionate stratified random sampling technique. The results indicated that households were technically inefficient in potato production by 44.7
Potato (Solanum tuberosum L.) is one of the most important crops worldwide; however, the chemical basis of flavor in native tubers remains poorly understood. This study evaluated the effects of tuber condition (with and without peel) and heat treatment (boiling and baking) on the physical properties, chemical composition, and sensory acceptability of three native potato varieties and one commercial variety (Chola), belonging to the phureja and tuberosum groups. Puña negra exhibited darker coloration, while Leona blanca, Chola, and Yema de huevo showed higher brightness and color saturation, which were preferred by panelists. Tubers processed with peel presented higher sensory acceptability, associated with increased total sugars, particularly in Yema de huevo. The peel contributed to structural integrity and nutrient retention, whereas its removal reduced glycoalkaloid content, improving safety. Thermal processing increased the energy value of potatoes compared to raw tubers, with effects influenced by peel presence. Overall, significant differences were observed among varieties, tuber condition, and heat treatment, with clear associations between chemical composition and sensory attributes.
Efficient water management and the use of biostimulants are emerging as sustainable approaches to enhance crop productivity under conditions of decreasing water availability. A 2-year field experiment was conducted during the rabi seasons of 2014–2015 and 2015–2016 at Banaras Hindu University, Varanasi, India, to evaluate the effects of irrigation schedules and amino acid biostimulants (AAB) on water use efficiency (WUE), residual soil moisture content (RSMC), growth and yield of potato (Solanum tuberosum L.). The experiment was laid out in a split-plot design with three irrigation schedules as main-plot treatments: three irrigations (I1), four irrigations (I2) and five irrigations (I3), and four AAB treatments as sub-plot treatments: control (A0), animal-based (Aa), plant-based (Ap), and mixed amino acid formulation (Am). Foliar sprays of AAB were applied three times during crop growth. Plant growth parameters, tuber yield, WUE and RSMC at different depths were recorded. The results revealed that irrigation schedules and AAB treatments significantly influenced potato growth, yield, WUE and soil moisture retention. The four-irrigation schedule (I2) produced the highest tuber yield (23.06 and 22.62 t ha−1) and WUE (115.05 and 93.16 kg ha−1 mm−1) during Trials 1 and 2, respectively. Among the biostimulants, the plant-based formulation (Ap) consistently recorded the highest yield (23.90 and 23.42 t ha−1) and WUE (119.25 and 96.48 kg ha−1 mm−1), followed by the mixed and animal-based formulations. The interaction treatment I2Ap resulted in the maximum yield (25.57 and 24.13 t ha−1) and WUE (127.57 and 99.41 kg ha−1 mm−1) in the respective years. Residual soil moisture content increased with irrigation frequency and AAB application, with the highest moisture retention observed under I3Ap in the 0–30 cm soil layer. Improved soil moisture availability enhanced nutrient uptake, plant growth and physiological processes, contributing to higher productivity and WUE. The study demonstrates that integrating a four-irrigation schedule with plant-based amino acid biostimulants is an effective strategy for improving potato yield, WUE and RSMC in Inceptisols of the Indo-Gangetic Plains, thereby supporting sustainable potato production under limited water resources. A graphical abstract illustrating the sustainable utilisation of organic waste for biostimulants production and its application to improve residual soil moisture and WUE.
Potato prices are shaped by local production cycles, interregional market co-movement, and recurrent seasonal fluctuations, making reliable forecasting difficult when only historical price records are available. This paper develops MarketFM, a price-only self-supervised foundation model for regional potato price forecasting. Instead of training a supervised predictor directly from limited labelled windows, MarketFM first learns a general representation of agricultural market dynamics from historical provincial price panels. The proposed framework tokenizes regional price histories into temporal patches, encodes both within-province evolution and cross-province co-movement through a factorised market encoder, and pretrains the encoder using masked price modeling, market-view contrastive learning, and frequency-temporal alignment. These self-supervised objectives encourage the model to recover missing price segments, learn perturbation-invariant market states, and preserve oscillatory price structures without relying on weather, logistics, inventory, policy, or other external variables. Forecasts are obtained by adapting the pretrained encoder with a lightweight horizon-aware decoder. Experiments on weekly potato prices from 25 Chinese provinces during 2012 to 2018 show that MarketFM achieves lower MAE, RMSE, and MAPE than CNN, LSTM, N-BEATS, Autoformer, and Informer on both one-step and four-step horizons. Component analysis further indicates that removing self-supervised pretraining, masked price modeling, market-view contrastive learning, or frequency-temporal alignment consistently weakens performance. The results suggest that price-only self-supervised market representation learning is an effective and practical strategy for agricultural price forecasting under limited data conditions.
