Drought Induced Metabolomics of Potato Leaves Highlight Metabolic Reprogramming and Promising Biomarkers for Smart Irrigation Advisories

Drought induced metabolomics of potato leaves highlight metabolic reprogramming and 1
promising biomarkers for smart irrigation advisories 2
Portia D Singh 1, Ramnath Nayak 1, Yvonne Dittrich2, Radoslaw Guzinski3,Yogesh Pant1 and 3
Shyam Kumar Masakapalli1* 4

1. School of Biosciences and Bioengineering, Indian Institute of Technology Mandi, Kamand, 5
   Himachal Pradesh-175075, India 6
2. IT University, Copenhagen, Rued Langgaards Vej 7, 2300 København, Denmark, 7
3. DHI A/S, Agern Alle 5, 2970 Hørsholm, Denmark 8
   Email: portia.iitmandi@gmail.com 9
   Email: ramnathnayak01@gmail.com 10
   Email: ydi@itu.dk 11
   Email: rmgu@dhigroup.com 12
   Email: yogeshp.iitmandi@gmail.com 13

• For correspondence: Email: shyam@iitmandi.ac.in 14
  Graphical abstract 15
  16
  Keyword: Drought, potato, metabolomics, early biomarkers, satellite imaging analysis 17
  18
  preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
  The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

Abstract 19
Smart irrigation management is essential for improving crop resilience under increasing 20
drought frequency driven by climate change. Although satellite-based remote sensing provides 21
valuable tools for monitoring crop water status at large spatial scales, its accuracy is often 22
limited in mountainous and heterogeneous agricultural landscapes. In this study, we 23
investigated drought -induced metabolic responses in potato ( Solanum tuberosum L.) to 24
identify biochemical biomarkers that could complement satellite-based irrigation advisories in 25
the mid-Himalayan region of India. A field experiment was conducted using a gradient of soil 26
moisture regimes corresponding to moderate (50% field capacity), critical (25% field capacity), 27
and extreme drought stress (5 -8% field capacity). Satellite-derived evapotranspiration-based 28
irrigation advisories were validated against in situ soil moisture measurements, revealing 29
discrepancies attributed to the inability of satellite estimates to capture actual water loss under 30
drought stress conditions, highlighting the need for additional ground -truth biomarkers across 31
heterogeneous field conditions. To capture plant -level physiological responses, untargeted 32
metabolite profiling of potato leaves was performed using gas chromatography –mass 33
spectrometry (GC -MS). Approximately fifty metabolites belonging to amino acids, organic 34
acids, sugars, and sugar alcohols were detected. Multivariate statistical analyses revealed 35
distinct metabolic signatures associated with progressive drought stress. Notably, accumulation 36
of proline, serine, isoleucine, sucrose, fructose, glucose, and polyols such as mannitol and myo-37
inositol reflected key metabolic reprogramming associated with osmoprotection, redox 38
homeostasis, and energy metabolism under drought conditions. Collectively, this ensemble of 39
stress-responsive metabolites represents a robust panel of drought stress biomarkers. As a proof 40
of concept, proline was validated as a qualitative biomarker of plant water status through a 41
rapid and cost -effective colorimetric biochemical assay, demonstrating its practical 42
applicability for field -level irrigation management. These findings demonstrate that 43
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

metabolomics-derived biomarkers can provide sensitive plant-level indicators of drought stress 44
that complement satellite -based monitoring systems. The integration of biochemical 45
diagnostics with remote sensing platforms offers a promising approach for improving drought 46
detection and developing low -cost, field -deployable tools for smart irrigation advisories in 47
heterogeneous agricultural landscapes. 48
1 Introduction 49
Potato (Solanum tuberosum L.) ranks among the most important staple food crops globally, 50
providing essential calories and nutrients to billions. However, its production is significantly 51
vulnerable to abiotic stresses, particularly drought, which severely limits tuber yield and quality 52
(Obidiegwu et al., 2015). The increasing frequency and severity of drought events due to 53
climate change necessitate the development of advanced, efficient methods for early stress 54
detection to enable timely irrigation management and mitigate crop losses ( Satapathy et al., 55
2024; Disasa et al., 2025). Traditional approaches for assessing water stress often rely on visual 56
symptoms, which typically manifest only after irreversible physiological damage and yield 57
potential reduction have occurred. While physiological measurements offer more sensitivity, 58
they can be labor-intensive and impractical for large-scale or rapid assessments (Jones; 2007). 59
The era of Smart Agriculture offers powerful alternatives by leveraging data -driven 60
technologies for precision crop management. Among these, satellite -based remote sensing 61
provides invaluable tools for monitoring crop conditions across extensive agricultural areas. 62
Techniques utilizing multispectral, hyperspectral, or thermal imagery can assess parameters 63
like vegetation indices (e.g., NDVI) , canopy temperature and evapotranspiration , which 64
correlate with plant water status and photosynthetic activity, offering a macro-scale perspective 65
on potential stress development (Ahmad et al., 2021). Although satellite remote sensing 66
provides valuable spatio -temporal coverage for monitoring vegetation dynamics, its reliance 67
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

on broad spectral indices limits the accurate identification of crop -specific stress responses. 68
Because these signals are indirect and influenced by multiple environmental and biological 69
factors, molecular or ground-based physiological measurements are necessary to validate and 70
accurately interpret stress status (Wen et al., 2021 and Co oper et al., 2025) . Metabolomics 71
offers a multivariate understanding of plant biochemical status and, when integrated with 72
satellite-based or high-throughput remote sensing, can enhance early detection and diagnosis 73
of crop diseases by linking molecular responses to observable stress patterns (Arbona et al., 74
2013; Terentev & Dolzhenko, 2023).Metabolic read-outs offer systems-level insights into plant 75
physiological adjustments to environmental conditions. Plants exhibit diverse phenotypic 76
plasticity under water deficit, marked by early sensing and signaling, hormonal activation 77
(primarily ABA), physiological and metabolic adjustments, structural adaptations, and 78
eventual recovery or senescence depending on stress duration. Under drought stress, potato 79
plants accumulate or deplete specific metabolites such as proline, soluble sugars, abscisic acid, 80
and various amino acids as part of their adaptive response (Gong et al., 2015; Vasquez-Robinet 81
et al., 2008). These molecules serve as sensitive biomarkers, enabling early and specific 82
detection of water deficit at the cellular level, often before physiological changes become 83
reliably detectable from space (Arbona et al., 2013; Obata & Fernie, 2012 ). Integrating the 84
broad spatio -temporal monitoring capabilities of satellite -based imaging with the high 85
sensitivity of metabolomic biomarkers provides a powerful, complementary approach for 86
managing drought stress in potato. By linking field -scale remote sensing with precise 87
biochemical indicators, this strategy enables more accurate, timely, and targeted interventions 88
to enhance crop resilience and productivity under water-limited conditions (Ahmad et al., 2021; 89
Obata & Fernie, 2012). 90
We hypothesize that metabolite -level responses in potato leaves provide early and sensitive 91
biomarkers of drought stress that can complement satellite -based irrigation advisories in 92
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

heterogeneous mountainous terrains. Therefore, in this study, we aimed to (i) demonstrate that 93
satellite-based drought advisories require complementary support from high-resolution ground 94
observations, particularly in hilly terrain with small and heterogenous agricultural fields, and 95
(ii) establish metabolic biomarkers of drought stress in potato ( Kufri Jyoti) by integrating soil 96
moisture gradients with GC -MS-based untargeted metabolite profiling. These findings 97
underscore the potential for developing low-cost, field-deployable biochemical assays for real-98
time drought diagnostics, thereby providing a complementary platform to remote sensing in 99
topographically complex terrains and enabling more precise, data-driven irrigation advisories. 100
2 Material and Method 101
2.1 Plant material 102
The study was conducted in March- April 2024 during the potato (Solanum tuberosum cv. Kufri 103
Jyoti) growing season in a field located in Sakaryar Village, Mandi, Himachal Pradesh, India 104
(Latitude: 31.7724520, Longitude: 77.0375349). Two experimental plots were systematically 105
prepared and fertilized to ensure adequate nutrient availability. Environmental data collection 106
included daily recordings of minimum and maximum temperatures (°C) and average relative 107
humidity (%) using a Tempnote TH32 Portable Temperature Data Logger. Soil moisture (%) 108
was measured daily using a Lutron PMS-714 Soil Moisture Meter. The soil moisture readings 109
were calibrated and correlated with field capacity (Supplementary Table 1). To complement 110
ground-based observations, daily evapotranspiration (ET) data for the specified field 111
coordinates were derived through satellite image analysis (Guzinski et al., 2020 & Guzinski et 112
al., 2021) ,with the objective of establishing correlations between satellite -derived and field -113
measured environmental variables for deriving irrigation advisories (Supplementary Table 1). 114
2.2 Experimental design for water stress assessment 115
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

The water stress experiment was conducted over two weeks in the field (March 2 2 - April 5, 116
2024) at two locations: Plot A (control, regularly irrigated) and Plot B (subjected to drought 117
stress). Drought conditions in Plot B were categorized into three soil moisture (SM) regimes 118
based on field capacity (FC), calculated from daily weight loss: (1) moderate stress (SM 50, 119
50% FC), (2) critical stress (SM 25, 25% FC), and (3) extreme stress (SM 0, 5-8% FC).Satellite 120
image analysis was performed using the Sen-ET modelling approach (Guzinski et al., 2020 & 121
Guzinski et al., 202 1), covering the experimental site. In this approach thermal observations 122
from Sentinel-3 satellites (daily overpass, 1 km spatial resolution) are used synergistically with 123
shortwave optical observations from Sentinel-2 satellites (overpass every 5 days, 20 m spatial 124
resolution), meteorological forcing and landcover map to drive a physically based actual 125
evapotranspiration (ET) model. The land cover map, used for ET model parameterization, was 126
adjusted to reflect the location of the potato fields to enhance the accuracy of 127
evapotranspiration and water deficit modelling. The output of the ET model were actual and 128
potential evapotranspiration maps (in mm/day) with 20 m by 20 m pixel size. The maps were 129
produced daily, however on days in which the potato field was obscured by clouds during the 130
satellite data acquisition, a gap -filling approach was used instead of the physical model 131
(Guzinski et al., 2021) . The gap-filling uses the last available clear -sky ET estimates and 132
accounts for changing meteorological conditions (including solar irradiance) but does not 133
consider soil drying or wetting during. To estimate water stress, water deficit was calculated, 134
defined as (potential ET – actual ET). 135
2.3 Proline content test 136
Proline quantification in potato leaf samples was performed using the RealGene ™ Proline 137
Content Estimation Kit, following the manufacturer’s recommended protocol. Approximately 138
100 mg of fresh leaf tissue was homogenized in 1 mL of extraction reagent and centrifuged at 139
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

10,000 × g for 10 minutes at 4°C. From the resulting supernatant, 500 µL was combined with 140
500 µL of Reagent 1 (glacial acetic acid) and 500 µL of Reagent 2, followed by incubation at 141
100°C for 60 minutes. After vortexing, the mixture was subjected to spectrophotometric 142
analysis, with absorbance measured at 520 nm. Proline concentration was determined using a 143
standard calibration curve generated from L -proline standards, and results were expressed as 144
micromoles per gram of fresh weight (µmol g⁻¹ FW). All measurements were conducted in 145
biological replicates (n = 4) to ensure statistical reliability. For comparison, L-proline standard 146
solutions ranging from 0 to 40 µg/mL were prepared. 147
2.4 Metabolite Profiling Workflow: Sample Preparation, Metabolite Extraction, and GC-148
MS Data Acquisition 149
To profile metabolites, lyophilized potato leaf tissue (~50 mg) was used for soluble metabolite 150
extraction following the protocol described by Lisec et al. , (2006). Each sample was 151
homogenized in 940 µL of extraction solvent comprising methanol:chloroform:water (3:1:1, 152
v/v/v) and supplemented with 60 µL of ribitol (0.2 mg/mL in H₂O) as an internal standard. 153
Samples were incubated at 70°C for 5 minutes in a thermomixer at 950 rpm, followed by 154
centrifugation at 13,000 × g for 10 minutes at room temperature. Approximately 50 µL of the 155
supernatant was transferred to a fresh microcentrifuge tube and vacuum dried. Dried extracts 156
were derivatized via methoxyamine hydrochloride (MeOX) and trimethylsilyl (TMS) reagent 157
treatment, as described by Masakapalli et al. , (2014). Samples were subsequently analyzed 158
using a GC -MS system (GC ALS -MS 5977B, Agilent Technologies) equipped with an HP -159
5ms capillary column (30 m × 250 μm × 0.25 μm). Metabolite identification was based on 160
comparison of mass ion fragments (m/z), retention times (RT), and characteristic identifier ions 161
with spectral data from the NIST 17 library. For statistical and multivariate data analysis, we 162
utilized MetaboAnalyst 6.0, accessible at https://www.metaboanalyst.ca/. Data pre-processing 163
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

was performed by normalization by the median, log transformation, and Pareto scaling. The 164
area of the internal standard, Ribitol, is used to obtain the relative proportions of the peaks, 165
leading to the calculation of fold changes of metabolite/ peak levels among the treatments. 166
3 Results and Discussion 167
3.1 Integration of satellite image analysis with ground-truth measurements on potato 168
field crop for enhanced drought monitoring in mountainous terrain 169
Daily environmental parameters for field experiments were recorded, including minimum and 170
maximum temperatures (°C), average relative humidity (%), and soil moisture content (%) 171
(Supplementary Table 2). Notably, discrepancies emerged between ground-based soil moisture 172
data and satellite -derived evapotranspiration data. While satellite -based advisories indicated 173
"watered field" conditions, ground measurements revealed a progressive decline in soil 174
moisture to 50 -25% of field capacity (Figure 1, Supplementary Table 1). Satellite advisories 175
continued signaling adequate moisture across drought-stressed plot until 2 April 2024, failing 176
to detect localized water stress under both mild (SM 50) and critical (SM 25) drought 177
conditions. Only after this date did satellite observations indicate irrigation requirements 178
(Supplementary Table 3). There are several potential sources for this inconsistency . Satellite 179
image analysis in mountainous regions like Himachal Pradesh is challenging due to complex 180
topography, variable land cover, frequent cloud cover, and small fragmented plots 181
compromising remotely sensed data accuracy ( Ma et al., 2025; Chen et al., 2023; Zhu et al., 182
2013; Zhu et al., 201 2). Specifically, during the experiment there was a gap in cloud -free 183
satellite observations between 25/03/2024 and 02/04/2024. Therefore, the transition from mild 184
drought to extreme drought for plot B (Figure 1) was not observed by satellites and instead the 185
gap-filled ET values, which do not consider soil drying, were used for satellite-based irrigation 186
advisory. As soon as satellite data become available, the advisory changed to “Required 187
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

Watered”. Additionally, advisories derived from 20×20 m spatial resolution data represented 188
mean values across multiple small fields within the same pixel. Sparse ground -truth data in 189
such terrains further hinders proper calibration and validation of satellite models (Brocca et al., 190
2011). Finally, the analysis relied on a generic vegetation response model to translate modeled 191
potential and actual evapotranspiration into plant water stress. However, as demonstrated by 192
our results, potato being a tuber crop exhibits a crop -specific water response function that 193
differs substantially from that of generic vegetation parameterizations. Consequently, satellite-194
based drought advisories in such agro-ecological settings should be complemented with high-195
resolution ground observations and crop -specific, plant -level diagnostics to account for 196
physiological and phenological differences, thereby improving the reliability and accuracy of 197
drought stress assessment. 198
3.2 Multivariate analysis reveals distinct metabolic fingerprints across progressive 199
drought stress levels 200
GC-MS-based metabolite profiling of the potato leaves detected approximately 50 metabolites 201
belonging to the classes of amino acids, fatty acids, organic acids, sugars, and sugar alcohols 202
(Supplementary Table 4). Multivariate analyses reveal ed clear metabolic variations among 203
control, and drought treatments of SM 50 (moderate stress), SM 25 (critical stress), and SM 0 204
(extreme stress) (Figure 2). PCA captured 74.75% of variance across PC1 and PC2, effectively 205
separating the treatment groups (Figure 2.A). Control samples are characterized by higher 206
glycine, glucopyranoside, pinitol, ethanolamine, and citric acid . Pentanedioic acid strongly 207
associates with SM 50, while mannobiose and serine contribute to its separation. SM 25 and 208
SM 0 exhibit elevated threonine, isoleucine, arbutin, erythrose, hydroxylamine, monopalmitin, 209
oxoproline, and mannobiose, indicating stress -related metabolic shifts. The VIP score plot 210
identified proline as the most discriminant metabolite, peaking under SM 25, suggesting a 211
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

strong adaptive response to moderate drought (Figure 2.B). Changes in amino acids (proline, 212
valine, tryptophan, serine, glycine, threonine) reflected osmoprotection, stress signaling, and 213
biosynthetic roles in drought-stressed plants (Obata & Fernie, 2012). Proline accumulation, in 214
particular, serves multifunctional roles in stress tolerance (Szabados & Savouré, 2010), while 215
branched-chain and aromatic amino acids participate in broader metabolic reprogramming 216
(Less & Galili, 2008). Variations in organic acids (citric, fumaric, succinic, citramalic) 217
reflected TCA cycle adjustments and energy production under drought stress (Araújo et al., 218
2012; Nunes-Nesi et al., 2010), consistent with central carbon-nitrogen metabolic regulation in 219
photosynthetic tissues. Pinitol variation tied to osmotic stress tolerance responses in potato 220
(Chourasia et al.,2021), functioning as a compatible solute and osmoprotectant. PLS-DA biplot 221
clearly separates groups, highlighting citric acid and pinitol in control, tryptophan in SM 50, 222
proline and valine in SM 25, and phytol in SM 0 (Figure 2.C). Hierarchical clustering reveals 223
tight within-group similarity with distinct treatment -based clustering (Figure 2.D). Notably, 224
SM 25 and SM 0 cluster closely, as do control and SM 50, suggesting shared metabolic 225
responses within pairs. These analyses confirm drought stress induce d metabolic 226
reprogramming, particularly affecting amino acid and organic acid pools reflective of adaptive 227
physiological responses. 228
3.3 Drought-induced amino acid accumulation in potato leaves is a key adaptive response 229
to declining soil moisture 230
Progressive drought stress triggered substantial metabolic reprogramming in potato leaves, 231
particularly marked by amino acid accumulation peaking under critical stress (SM 25), with 232
significant increases in proline, valine, serine, threonine, tryptophan, and phenylalanine (Figure 233
3). Proline showed the most dramatic rise, functioning as a central osmoprotectant and stress 234
marker (Szabados & Savouré, 2010; Singh et al., 2015). Amino acid accumulation served 235
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

multiple adaptive functions: supporting osmotic adjustment by lowering cellular water 236
potential to maintain turgor (Chaves et al., 2003), accumulating from drought-induced growth 237
inhibition and enhanced protein degradation (Hildebrandt et al., 2015), and providing ROS 238
scavenging and membrane stabilization. Aromatic amino acids like phenylalanine and 239
tryptophan serve as precursors for defense compounds via the shikimate pathway (Galili et al., 240
2016; Tzin & Galili, 2010), while accumulated amino acids may represent temporary nitrogen 241
storage under stress conditions (Less & Galili, 2008). Distinct response patterns including early 242
aspartic acid peak at SM 50, continued isoleucine increase to SM 0, and glycine decrease 243
highlighted the complexity of amino acid metabolic regulation in potato leaves under 244
progressive drought stress. These differential accumulation patterns likely reflect amino acid -245
specific pathway regulation, differential stress signal transduction, and dynamic shifts in 246
metabolic flux priorities as drought stress intensified (Galili et al., 2016; Less & Galili, 2008). 247
Such coordinated but differentiated amino acid responses represent a critical adaptive strategy 248
in potato, integrating osmotic regulation, resource management, cellular protection, and 249
signaling pathways to enhance plant survival during water scarcity. 250
3.4 Alterations within nitrogen metabolism and its pathways to varying drought stress 251
intensities 252
Under drought stress in potato leaves, oxoproline accumulated significantly under critical -to-253
extreme conditions (SM 25 & SM 0) (Supplementary Figure S1.A), likely reflecting increased 254
proline synthesis and cycling for osmoprotection, consistent with proline -glutamate pathway 255
responses characterized across plant species (Verbruggen & Hermans, 2008; Szabados & 256
Savouré, 2010). Notably, GABA levels remained stable across all drought severities, despite 257
GABA being commonly induced by abiotic stress (Ramesh et al., 2017). This unexpected 258
stability in potato leaves may indicate high metabolic turnover via the GABA shunt (Michaeli 259
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

& Fromm, 2015) or reflect potato -specific regulation of nitrogen metabolism under water 260
deficit. Hydroxylamine displayed a biphasic response, declining sharply under mild stress (SM 261
50) then spiking at critical stress (SM 25), indicating complex, stress -intensity- dependent 262
shifts in nitrogen metabolism. As hydroxylamine participates in nitric oxide metabolism and 263
nitrogen redox pathways (Astier et al., 2018; Gupta et al., 2011; Mur et al., 2013), its dynamic 264
accumulation pattern in potato leaves likely reflects disrupted nitrogen flux or altered 265
enzymatic activity under progressive drought conditions. These nitrogen metabolites 266
collectively highlight specific, stress -intensity-dependent metabolic reprogramming in potato 267
leaves responding to water scarcity. 268
3.5 Alterations in fatty acids metabolic pathways of potato leaves in response to drought 269
stress 270
Significant alterations in fatty acid metabolism and signaling pathways occurred in potato 271
leaves in response to declining soil moisture. Saturated fatty acids (palmitic, stearic) and their 272
monoacylglycerols (monopalmitin, glycerol monostearate) decreased under m oderate stress 273
(SM 50) but peaked under critical -to-extreme stress (SM 25 ,SM 0 ) (Supplementary Figure 274
S1.B). This accumulation likely reflected drought -induced membrane lipid remodeling, with 275
increased saturated fatty acid proportions decreasing membrane fluidity, an adaptation 276
commonly observed under abiotic stress in leaf tissues (Upchurch, 2008; Gigon et al., 2004). 277
Conversely, the polyunsaturated fatty acid linolenic acid (C18:3) substantially decreased under 278
all drought levels (SM 50, SM 25, SM 0), potentially due to increased lipid peroxidation from 279
oxidative stress (Noctor et al., 2014), selective degradation of chloroplast membranes where 280
linolenic acid is highly enriched, or diversion into jasmonic acid biosynthesis (Wasternack & 281
Hause, 2013). Additionally, progressive accumulation of the biogenic amine octopamine under 282
critical (SM 25) and extreme (SM 0) drought conditions suggested activation of specific stress 283
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

signaling cascades in potato leaves. Catecholamines including octopamine and tyramine 284
modulate stress responses in potato through effects on metabolism and physiology (Swiedrych 285
et al., 2004; Kulma & Szopa, 2007), potentially influencing the observed metabolic 286
adjustments under water deficit. Together, these coordinated changes in saturated lipids, 287
unsaturated fatty acid depletion, and signaling amine accumulation highlight the intricate 288
biochemical adjustments in lipid metabolism and stress signaling employed by potato leaves 289
to cope with water deficit. 290
3.6 Metabolic reprogramming of central carbon metabolism and sugar accumulation 291
under progressive drought stress 292
Drought stress significantly perturbs central carbon metabolism, revealing nuanced metabolic 293
reprogramming rather than uniform responses (Supplementary Figure S2). TCA cycle 294
intermediates show varied patterns: significant reductions in succinic acid (SM 25, SM 0) and 295
malic acid (SM 0) likely result from reduced respiratory demand (Atkin & Macherel, 2009), 296
limited CO₂ uptake affecting acetyl -CoA supply (Lawlor & Cornic, 2002), and diversion of 297
carbon skeletons such as oxaloacetate and α-ketoglutarate toward synthesis of osmoprotective 298
amino acids (Less & Galili, 2008; Araújo et al., 2012; Hildebrandt et al., 2015). Citric acid 299
decreased under mild stress (SM 50) but recovered under stronger stress, potentially reflecting 300
its multifunctional roles in cytosolic biosynthesis, cellular pH homeostasis, and metabolic 301
regulation (Igamberdiev & Eprintsev, 2016; Sweetlove et al., 2010). Fumaric acid and 302
citramalic acid remained stable, suggesting tight metabolic regulation or lower sensitivity to 303
drought stress. 304
Carbohydrate metabolism in potato leaves showed multifaceted shifts under drought stress 305
(Figure 4) . Under mild stress (SM 50), sucrose, glucose, and fructose accumulated 306
significantly, consistent with source-sink imbalance where growth inhibition outpaced reduced 307
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

photosynthetic carbon fixation (Lawlor & Cornic, 2002). As stress intensified (SM 25, SM 0), 308
these sugars declined due to increased conversion into protective compounds including amino 309
acids and polyols ( Hare et al., 1998 ), elevated energy demands for cellular maintenance, and 310
photosynthetic feedback inhibition (Paul & Foyer, 2001). Under moderate -to-severe stress, 311
accumulation of myoinositol, mannobiose, and erythrose reflected active carbohydrate 312
metabolic reprogramming. Myoinositol and related polyols function as compatible solutes in 313
drought-stressed potato leaves (Legay et al., 2011), serving as precursors for osmoprotectants 314
like pinitol and galactinol while contributing to membrane stabilization and signaling (Valluru 315
& Van den Ende, 2011). Polyols including mannitol contribute to ROS scavenging and 316
oxidative protection (Smirnoff & Cumbes, 1989; Hare et al., 1998), processes critical for 317
maintaining photosynthetic capacity under water deficit (Obidiegwu et al., 2015). Overall, 318
these coordinated shifts highlight the metabolic flexibility of potato leaves, implementing 319
strategic carbohydrate-to-polyol conversion supporting osmoprotection, energy balance, and 320
stress adaptation (Legay et al., 2011; Pinheiro & Chaves, 2011). 321
3.7 Drought stress induces metabolic shifts enhancing osmoprotection, energy balance, 322
and stress resilience in plants 323
Drought stress induced extensive metabolic reprogramming in potato leaves to maintain 324
cellular homeostasis and enhance stress tolerance (Figure 5). Primary affected pathways 325
included glycolysis, photorespiration, the shikimate pathway, TCA cycle, GS/GOGAT cycle, 326
branched-chain amino acid (BCAA) biosynthesis, and sugar metabolism, consistent with 327
drought responses observed across potato genotypes (Sprenger et al., 2016; Vasquez -Robinet 328
et al., 2008). Drought -induced accumulation of sucrose, glucose, fructose, pinitol, mannitol, 329
and galactaric acid reflected altered osmoprotection and energy management (Evers et al., 330
2010; Legay et al., 2011). Elevated sucrose, mannitol, and erythritol levels indicated enhanced 331
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

compatible solute accumulation to stabilize cellular structures and scavenge reactive oxygen 332
species (Loescher, 1987; Hare et al., 1998). Photorespiration intensified under drought, 333
evidenced by increased glycine -serine cycling, functioning in stress tolerance and redox 334
balancing (Wingler et al., 2000). Shikimate pathway flux changes drove biosynthesis of 335
phenylalanine, tryptophan, and phenolic compounds including arbutin, crucial for secondary 336
metabolite production and defense signaling (Tzin & Galili, 2010; Sprenger et al., 2018). 337
Enhanced BCAA and TCA-derived amino acid biosynthesis suggested prioritization of protein 338
and osmolyte synthesis from altered TCA cycle carbohydrate flows (Hildebrandt et al., 2015; 339
Araújo et al., 2012). Upregulated fatty acid biosynthesis (palmitic, stearic, monopalmitin, 340
linolenic acids) reinforced membrane stability (Upchurch, 2008). The GS/GOGAT cycle 341
showed increased glutamate, glutamine, and glutathione production for nitrogen assimilation 342
and ROS protection (Less & Galili, 2008). TCA cycle intermediates (citrate, succinate, 343
fumarate, malate, oxaloacetate) displayed varied accumulation patterns, reflecting adaptive 344
changes in respiratory energy production and redox homeostasis (Atkin & Macherel, 2009; 345
Sweetlove et al., 2010). Together, these metabolic shifts demonstrated a coordinated drought 346
response in potato leaves, maintaining redox balance, cellular integrity, and adaptive capacity 347
(Obidiegwu et al., 2015). 348
3.8 Quantification of proline using an enzymatic and colorimetric assay kit 349
Gas chromatography -mass spectrometry (GC -MS)-based metabolic profiling is a powerful 350
approach for elucidating plant responses to drought stress. However, GC -MS remains largely 351
inaccessible to breeders and farmers unfamiliar with metabolomics. To bridge this gap, a 352
simple, rapid, and cost -effective colorimetric assay for quantifying proline was evaluated to 353
facilitate screening of drought -tolerant cultivars under field conditions (Figure 6.A). A clear 354
physiological response to water deficit was observed (Figure 6.B), with proline accumulation 355
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

increasing markedly as soil moisture decreased, consistent with its established role as a key 356
osmoprotectant in plant drought tolerance. In field trials, proline content under critical drought 357
stress (SM 25) was approximately two-fold higher than in well-watered controls (Figure 6.C). 358
The distinct colorimetric changes associated with proline accumulation can be leveraged in two 359
directions: (i) training machine learning models for integration into mobile -based diagnostic 360
platforms, and (ii) developing low -cost, field -deployable biochemical assays for smart 361
irrigation advisories (Figure 6.D). Notably, under extreme drought (SM0), proline levels 362
remained comparable to controls, potentially indicating metabolic exhaustion or adaptive 363
regulation where severe stress restricts further synthesis and accumulation (Bhaskar et al., 364
2015). While proline accumulation is not entirely drought -specific and may increase under 365
other abiotic stresses such as salinity or extreme temperatures, and colorimetric assays can be 366
influenced by interfering compounds, this assay remains a rapid and practical tool for field -367
level screening. These findings highlight proline's reliability as a biochemical marker for 368
drought stress and support its utility in screening drought -responsive cultivars (Szabados & 369
Savouré, 2010 ; Singh et al., 2015 ). Furthermore, the broader ensemble of drought stress -370
responsive metabolites identified in this study shows a promising multi -biomarker platform 371
that can be similarly employed for high -throughput physiological screening and precision 372
drought stress assessment across diverse genotypes and field conditions. 373
4 Conclusion 374
In conclusion, while satellite imagery provides valuable broad spati o-temporal coverage and 375
performs reliably in plains, our findings reveal significant limitations in mountainous regions 376
where complex topography, cloud cover, small and heterogenous agricultural fields and sparse 377
calibration data hinder accurate drought detection. This study demonstrates that metabolomic-378
based biomarker analysis offers a robust complementary approach for drought stress 379
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

management in potato crops, particularly in challenging terrains. Through GC-MS profiling, 380
we identified a panel of early -stage (serine, isoleucine, pinitol) and late -stage (proline, 381
myoinositol, valine, sucrose, fructose, glucose) metabolic biomarkers reflecting critical 382
physiological processes such as osmotic adjustment and oxidative stress defence. Proline 383
emerged as a particularly practical biomarker due to its consistent accumulation and simple 384
colorimetric quantification. The proof -of-concept proline assay demonstrated field -level 385
applicability through (i) integration with machine learning for mobile diagnostic platforms, and 386
(ii) development of low-cost biochemical assays for timely irrigation advisories. 387
Overall, this study demonstrates the metabolomic biomarkers as a complementary strategy for 388
drought detection, bridging plant physiology with practical agricultural tools. By enabling 389
earlier and more precise monitoring in topographically complex regions where satellite 390
advisories prove insufficient, these approaches can support optimized irrigation, improved 391
water-use efficiency, and enhanced resilience of smallholder farming systems in water -scarce 392
environments. Further, a better understanding of plant metabolic can be used to better calibrate 393
satellite image analysis to the specific crop, even for plains. 394
Acknowledgment 395
The authors are thankful to the SaIAFarm Team for all the data access. PDS acknowledge Mr 396
Jagdish Chand (field staff) for continuous monitoring of experimental field, SKM acknowledge 397
DST for funding support under INT_Denmark_P-4_2020 (G). PDS acknowledges project and 398
PhD fellowships from the Ministry of Education and DST, Govt of India 399
Contributions 400
PDS - conceptualization, data curation, formal analysis, validation, visualization, writing-401
original draft , RN - methodology, writing-review and editing, YD - funding acquisitio n, 402
writing-review and editing , RG - satellite image analysis, writing - review and editing, YP- 403
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

methodology, writing -review and editing , SKM - conceptualization, f unding acquisition , 404
supervision ,writing-original draft, review and editing and PDS and SKM agree to be 405
accountable for all aspects of the work. 406
Funding: SaIAFarm project (Sustainable Irrigation Advisories for Mid -Himalayan Farmers 407
using Smart Satellite Image Analytics) DST, Govt of India funded reference No 408
INT_Denmark_P-4_2020 (G) 409
Declarations of interest 410
Ethical approval: The study complied with the ethical standards. 411
Competing interests: The authors declare no competing interests. 412
Data availability: 413
The datasets generated during and/or analysed during the current study are available from the 414
corresponding author on reasonable request. 415
References 416

1. Ahmad, U., Alvino, A., & Marino, S. (2021). A review of crop water stress assessment using 417
   remote sensing. Remote Sensing, 13(20), 4155. https://doi.org/10.3390/rs13204155 418

2. Araújo, W. L., Nunes -Nesi, A., Nikoloski, Z., Sweetlove, L. J., & Fernie, A. R. (2012). 419
   Metabolic control and regulation of the tricarboxylic acid cycle in photosynthetic and 420
   heterotrophic plant tissues. Plant, Cell & Environment , 35(1), 1 -21. 421
   https://doi.org/10.1111/j.1365-3040.2011.02332.x 422

3. Arbona, V., Manzi, M., de Ollas, C., & Gómez-Cadenas, A. (2013). Metabolomics as a tool to 423
   investigate abiotic stress tolerance in plants. International Journal of Molecular Sciences , 424
   14(3), 4885–4911. https://doi.org/10.3390/ijms14034885 425

4. Astier, J., Gross, I., & Durner, J. (2018). Nitric oxide production in plants: an update. Journal 426
   of Experimental Botany, 69(14), 3401-3411. https://doi.org/10.1093/jxb/erx420 427

5. Atkin, O. K., & Macherel, D. (2009). The crucial role of plant mitochondria in orchestrating 428
   drought tolerance. Annals of Botany, 103(4), 581-597. https://doi.org/10.1093/aob/mcn094 429
   preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
   The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

6. Bhaskara, G. B., Yang, T. H., & Verslues, P. E. (2015). Dynamic proline metabolism: 430
   importance and regulation in water limited environments. Frontiers in Plant Science , 6, 484. 431
   https://doi.org/10.3389/fpls.2015.00484 432

7. Brocca, L., Hasenauer, S., Lacava, T., Melone, F., Moramarco, T., Wagner, W., ... & Bittelli, 433
   M. (2011). Soil moisture estimation through ASCAT and AMSR -E sensors: An 434
   intercomparison and validation study across Europe. Remote Sensing of Environment, 115(12), 435
   3390-3408.https://doi.org/10.1016/j.rse.2011.08.003 436

8. Chourasia, K. N., Lal, M. K., Tiwari, R. K., Dev, D., Kardile, H. B., Patil, V. U., ... & Pramanik, 437
   D. (2021). Salinity stress in potato: Understanding physiological, biochemical and molecular 438
   responses. Life, 11(6), 545. https://doi.org/10.3390/life11060545 439

9. Chaves, M. M., Maroco, J. P., & Pereira, J. S. (2003). Understanding plant responses to drought 440
   from genes to the whole plant. Functional Plant Biology, 30 (3), 239 -264. 441
   https://doi.org/10.1071/FP02076 442

10. Chen, R., Yin, G., Zhao, W., Yan, K., Wu, S., Hao, D., & Liu, G. (2023). Topographic 443
    correction of optical remote sensing images in mountainous areas: A systematic review. IEEE 444
    Geoscience and Remote Sensing Magazine , 11(4), 125 -145. 445
    https://doi.org/10.1016/j.isprsjprs.2025.05.027 446

11. Cooper, J., Propst, K., & Hirsch, C. D. (2025). Current methods and future needs for visible 447
    and non -visible detection of plant stress responses. Frontiers in Plant Science , 16, 448

12. 10.3389/fpls.2025.1585413 449

13. Disasa, K. N., Yan, H., Zhang, J., Wang, G., Zhang, C., Zhang, D., ... & Hao, B. (2025). 450
    Comprehensive review of drought characteristics and intensification under climate change: 451
    implications for agriculture and water resources. Journal of Hydrology , 134571. 452
    https://doi.org/10.1016/j.jhydrol.2025.134571 453

14. Evers, D., Lefevre, I., Legay, S., Lamoureux, D., Hausman, J. F., Rosales, R. O. G., ... & 454
    Schafleitner, R. (2010). Identification of drought -responsive compounds in potato through a 455
    combined transcriptomic and targeted metabolite approach. Journal of experimental 456
    botany, 61(9), 2327-2343. https://doi.org/10.1093/jxb/erq060 457

15. Galili, G., Amir, R., & Fernie, A. R. (2016). The regulation of essential amino acid synthesis 458
    and accumulation in plants. Annual review of plant biology , 67(1), 153 -178. 459
    https://doi.org/10.1146/annurev-arplant-043015-112213 460

16. Gigon, A., Matos, A. R., Laffray, D., Zuily -Fodil, Y., & Pham -Thi, A. T. (2004). Effect of 461
    drought stress on lipid metabolism in the leaves of Arabidopsis thaliana (ecotype Columbia). 462
    Annals of Botany, 94(3), 345-351.https://doi.org/10.1093/aob/mch150 463

17. Gong, L., Zhang, H., Gan, X., Zhang, L., Chen, Y., Nie, F., ... & Song, Y. (2015). 464
    Transcriptome profiling of the potato (Solanum tuberosum L.) plant under drought stress and 465
    preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
    The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

water-stimulus conditions. PloS one, 10(5), 466
e0128041.https://doi.org/10.1371/journal.pone.0128041 467
17. Gupta, K. J., Fernie, A. R., Kaiser, W. M., & van Dongen, J. T. (2011). On the origins of nitric 468
oxide. Trends in Plant Science, 16(3), 160-168.https://doi.org/10.1016/j.tplants.2010.11.007 469
18. Guzinski, R., Nieto, H., Sanchez, J. M., Lopez-Urrea, R., Boujnah, D. M., & Boulet, G. (2021). 470
Utility of copernicus -based inputs for actual evapotranspiration modeling in support of 471
sustainable water use in agriculture. IEEE Journal of Selected Topics in Applied Earth 472
Observations and Remote Sensing , 14, 11466 -11484. 473
https://doi.org/10.1109/JSTARS.2021.3122573 474
19. Guzinski, R., Nieto, H., Sandholt, I., & Karamitilios, G. (2020). Modelling high -resolution 475
actual evapotranspiration through Sentinel -2 and Sentinel -3 data fusion. Remote 476
Sensing, 12(9), 1433. https://doi.org/10.3390/rs12091433 477
20. Hare, P. D., Cress, W. A., & Van Staden, J. (1998). Dissecting the roles of osmolyte 478
accumulation during stress. Plant, Cell & Environment, 21 (6), 535 -479
553.https://doi.org/10.1046/j.1365-3040.1998.00309.x 480
21. Hildebrandt, T. M., Nesi, A. N., Araújo, W. L., & Braun, H. P. (2015). Amino acid catabolism 481
in plants. Molecular plant, 8(11), 1563-1579.https://doi.org/10.1016/j.molp.2015.09.005 482
22. Igamberdiev, A. U., & Eprintsev, A. T. (2016). Organic acids: the pools of fixed carbon 483
involved in redox regulation and energy balance in higher plants. Frontiers in Plant Science, 484
7, 1042. https://doi.org/10.3389/fpls.2016.01042 485
23. Jones, H. G. (2007). Monitoring plant and soil water status: established and emerging methods 486
revisited and their relevance to studies of drought tolerance. Journal of Experimental Botany, 487
58(2), 119–130. https://doi.org/10.1093/jxb/erl118 488
24. Kulma, A., & Szopa, J. (2007). Catecholamines are active compounds in plants. Plant 489
Science, 172(3), 433-440.https://doi.org/10.1016/j.plantsci.2006.10.013 490
25. Lawlor, D. W., & Cornic, G. (2002). Photosynthetic carbon assimilation and associated 491
metabolism in relation to water deficits in higher plants. Plant, Cell & Environment, 25 (2), 492
275-294.https://doi.org/10.1046/j.0016-8025.2001.00814.x 493
Legay, S., Lefèvre, I., Lamoureux, D., Barreda, C., Luz, R. T., Gutierrez, R., ... & Schafleitner, 494
R. (2011). Carbohydrate metabolism and cell protection mechanisms differentiate drought 495
tolerance and sensitivity in advanced potato clones (Solanum tuberosum L.). Functional & 496
integrative genomics, 11(2), 275-291. https://doi.org/10.1007/s10142-010-0206-z 497
26. Less, H., & Galili, G. (2008). Principal transcriptional programs regulating plant amino acid 498
metabolism in response to abiotic stresses. Plant Physiology, 147 (1), 316 -499
330.https://doi.org/10.1104/pp.108.115733 500
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

27. Lisec, J., Schauer, N., Kopka, J., Willmitzer, L., & Fernie, A. R. (2006). Gas chromatography 501
    mass spectrometry –based metabolite profiling in plants. Nature protocols , 1(1), 387 -396. 502
    https://doi.org/10.1038/nprot.2006.59 503
    Loescher, W. H. (1987). Physiology and metabolism of sugar alcohols in higher 504
    plants. Physiologia Plantarum, 70(3).https://doi.org/10.1111/j.1399-3054.1987.tb02857.x 505
    Ma, Y., Liang, S., Peng, W., He, T., Ma, H., Chen, Y., ... & Guan, S. (2025). A universal 506
    physically-based topographic correction framework for high -resolution optical satellite 507
    data. ISPRS Journal of Photogrammetry and Remote Sensing , 227, 459 -480. 508
    https://doi.org/10.1016/j.isprsjprs.2025.05.027 509

28. Masakapalli, S. K., Bryant, F. M., Kruger, N. J., & Ratcliffe, R. G. (2014). The metabolic flux 510
    phenotype of heterotrophic Arabidopsis cells reveals a flexible balance between the cytosolic 511
    and plastidic contributions to carbohydrate oxidation in response to phosphate limitation. The 512
    Plant Journal, 78(6), 964-977. https://doi.org/10.1111/tpj.12522 513

29. Michaeli, S., & Fromm, H. (2015). Closing the loop on the GABA shunt in plants: are GABA 514
    metabolism and signaling entwined? Frontiers in Plant Science, 6 , 419. 515
    https://doi.org/10.3389/fpls.2015.0041914 516

30. Mur, L. A., Mandon, J., Persijn, S., Cristescu, S. M., Moshkov, I. E., Novikova, G. V., ... & 517
    Harren, F. J. (2013). Nitric oxide in plants: an assessment of the current state of knowledge. 518
    AoB Plants, 5, pls052. https://doi.org/10.1093/aobpla/pls052 519

31. Noctor, G., Mhamdi, A., & Foyer, C. H. (2014). The roles of reactive oxygen metabolism in 520
    drought: not so cut and dried. Plant physiology , 164(4), 1636 -521
    1648.https://doi.org/10.1104/pp.113.233478 522

32. Nunes-Nesi, A., Fernie, A. R., & Stitt, M. (2010). Metabolic and signaling aspects 523
    underpinning the regulation of plant carbon nitrogen interactions. Molecular plant, 3(6), 973-524
    996.https://doi.org/10.1093/mp/ssq049 525

33. Obata, T., & Fernie, A. R. (2012). The use of metabolomics to dissect plant responses to abiotic 526
    stresses. Cellular and Molecular Life Sciences , 69(19), 3225 –3243. 527
    https://doi.org/10.1007/s00018-012-1091-5 528

34. Obidiegwu, J. E., Bryan, G. J., Jones, H. G., & Prashar, A. (2015). Coping with drought: stress 529
    and adaptive responses in potato and perspectives for improvement. Frontiers in plant 530
    science, 6, 542.https://doi.org/10.3389/fpls.2015.00542 531

35. Paul, M. J., & Foyer, C. H. (2001). Sink regulation of photosynthesis. Journal of Experimental 532
    Botany, 52(360), 1383-1400.https://doi.org/10.1093/jexbot/52.360.1383 533

36. Pinheiro, C., & Chaves, M. M. (2011). Photosynthesis and drought: can we make metabolic 534
    connections from available data?. Journal of experimental botany , 62(3), 869 -535
    882.https://doi.org/10.1093/jxb/erq340 536
    preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
    The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

37. Ramesh, S. A., Tyerman, S. D., Gilliham, M., & Xu, B. (2017). γ-Aminobutyric acid (GABA) 537
    signalling in plants. Cellular and Molecular Life Sciences , 74(9), 1577 -1603. 538
    https://doi.org/10.1007/s00018-016-2415-7 539

38. Satapathy, T., Dietrich, J., & Ramadas, M. (2024). Agricultural drought monitoring and early 540
    warning at the regional scale using a remote sensing -based combined index. Environmental 541
    Monitoring and Assessment, 196(11), 1132. https://doi.org/10.1007/s10661-024-13265-y 542

39. Singh, M., Kumar, J., Singh, S., Singh, V. P., & Prasad, S. M. (2015). Roles of osmoprotectants 543
    in improving salinity and drought tolerance in plants: a review. Reviews in environmental 544
    science and bio/technology, 14(3), 407-426. https://doi.org/10.1007/s11157-015-9372-8 545

40. Smirnoff, N., & Cumbes, Q. J. (1989). Hydroxyl radical scavenging activity of compatible 546
    solutes. Phytochemistry, 28(4), 1057-1060.https://doi.org/10.1016/0031-9422(89)80182-7 547

41. Sprenger, H., Erban, A., Seddig, S., Rudack, K., Thalhammer, A., Le, M. Q., ... & Hincha, D. 548
    K. (2018). Metabolite and transcript markers for the prediction of potato drought 549
    tolerance. Plant biotechnology journal, 16(4), 939-950. https://doi.org/10.1111/pbi.12840 550

42. Sprenger, H., Kurowsky, C., Horn, R., Erban, A., Seddig, S., Rudack, K., ... & Kopka, J. (2016). 551
    The drought response of potato reference cultivars with contrasting tolerance. Plant, cell & 552
    environment, 39(11), 2370-2389. https://doi.org/10.1111/pce.12780 553

43. Sweetlove, L. J., Beard, K. F., Nunes -Nesi, A., Fernie, A. R., & Ratcliffe, R. G. (2010). Not 554
    just a circle: flux modes in the plant TCA cycle. Trends in plant science , 15(8), 462 -470. 555
    https://doi.org/10.1016/j.tplants.2010.05.006 556

44. Świędrych, A., Lorenc -Kukuła, K., Skirycz, A., & Szopa, J. (2004). The catecholamine 557
    biosynthesis route in potato is affected by stress. Plant Physiology and Biochemistry, 42(7-8), 558
    593-600. https://doi.org/10.1016/j.plaphy.2004.07.002 559

45. Szabados, L., & Savouré, A. (2010). Proline: a multifunctional amino acid. Trends in Plant 560
    Science, 15(2), 89–97.https://doi.org/10.1016/j.tplants.2009.11.009 561

46. Terentev, A., & Dolzhenko, V. (2023). Can metabolomic approaches become a tool for 562
    improving early plant disease detection and diagnosis with modern remote sensing methods? 563
    A review. Sensors, 23(12), 5366. https://doi.org/10.3390/s23125366 564

47. Tzin, V., & Galili, G. (2010). The biosynthetic pathways for shikimate and aromatic amino 565
    acids in Arabidopsis thaliana. The Arabidopsis book/American Society of Plant Biologists , 8, 566
    e0132.https://doi.org/10.1199/tab.0132 567

48. Upchurch, R. G. (2008). Fatty acid unsaturation, mobilization, and regulation in the response 568
    of plants to stress. Biotechnology letters , 30, 967 -977.https://doi.org/10.1007/s10529-008-569
    9639-z 570
    preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
    The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

49. Valluru, R., & Van den Ende, W. (2011). Myo-inositol and beyond–emerging networks under 571
    stress. Plant Science, 181(4), 387-400. https://doi.org/10.1016/j.plantsci.2011.07.009 572

50. Vasquez-Robinet, C., Mane, S. P., Ulanov, A. V., Watkinson, J. I., Stromberg, V. K., De 573
    Koeyer, D., ... & Grene, R. (2008). Physiological and molecular adaptations to drought in 574
    Andean potato genotypes. Journal of experimental botany , 59(8), 2109 -2123. 575
    https://doi.org/10.1093/jxb/ern073 576

51. Verbruggen, N., & Hermans, C. (2008). Proline accumulation in plants: a review. Amino Acids, 577
    35(4), 753-759. https://doi.org/10.1007/s00726-008-0061-6 578

52. Wasternack, C., & Hause, B. (2013). Jasmonates: biosynthesis, perception, signal transduction 579
    and action in plant stress response, growth and development. An update to the 2007 review in 580
    Annals of Botany. Annals of botany, 111(6), 1021-1058. https://doi.org/10.1093/aob/mct067 581

53. Wen, W., Timmermans, J., Chen, Q., & van Bodegom, P. M. (2020). A review of remote 582
    sensing challenges for food security with respect to salinity and drought threats. Remote 583
    Sensing, 13(1), 6. https://doi.org/10.3390/rs13010006 584

54. Wingler, A., Lea, P. J., Quick, W. P., & Leegood, R. C. (2000). Photorespiration: metabolic 585
    pathways and their role in stress protection. Philosophical Transactions of the Royal Society of 586
    London. Series B: Biological Sciences , 355(1402), 1517 -587
    1529.https://doi.org/10.1098/rstb.2000.0712 588

55. Zhu, X., Chen, J., Gao, F., Chen, X., & Masek, J. G. (2013). An enhanced spatial and temporal 589
    adaptive reflectance fusion model for complex heterogeneous regions. Remote Sensing of 590
    Environment, 114(11), 2610–2623. https://doi.org/10.1016/j.rse.2010.05.032 591

56. Zhu, Z., & Woodcock, C. E. (2012). Object-based cloud and cloud shadow detection in Landsat 592
    imagery. Remote sensing of environment , 118, 83 -94 593
    https://doi.org/10.1016/j.isprsjprs.2025.05.027.594
    preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
    The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

595
Figure 1: Validation of satellite data and actual soil moisture level under different drought regimes. Soil moisture reading from sensors and 596
ET data advisories from March 22, 2024, to April 5, 2024.The green bar showed the soil moisture value at control field, orange bars represent soil 597
moisture values in drought field, the red text (watered /required water) represent the irrigation advisories based on ET values derived from satellite 598
image analysis. Notably, the satellite-based advisories failed to detect localized water stress in drought fields until 2 April 2024, despite declining 599
in-situ soil moisture levels (SM 50 and SM 25).600
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

601
Figure 2: Metabolites captured in potato leaves under drought treatment showed varied 602
metabolomes. (A) Principal component analysis (PCA) plot showing the separation of 603
different treatments based on metabolite profiles. Each treatment group (Control, SM 50, SM 604
25, SM 0) is represented by an ellipsoid, and key metabolites contributing to the variation are 605
labeled. (B) Variable Importance in Projection (VIP) scores of metabolites identified through 606
partial least squares (PLS) analysis, ranked by their contribution to the class separation. Colors 607
indicate the relative abundance of metabolites across different treatment groups. (C)The PLS -608
DA biplot for component 1 (39.7%) and component 2 (29.9%), showing the distribution of 609
metabolites in relation to experimental treatments. Significant metabolites contributing to each 610
group are highlighted aligning with the VIP analysis results. (D) Hierarchical clustering 611
dendrogram based on metabolite profiles, showing the clustering of treatment groups. The 612
dendrogram indicates similarities and differences in metabolic profiles between the control and 613
treatment groups (SM 50, SM 25, SM 0). 614
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

615
616
617
Figure 3: Drought exposure modulates amino acid metabolism in potato leaves. The levels 618
of amino acids (aspartic acid, tryptophan and phenylalanine ) are altered during SM 50 619
treatment. Higher increment of other amino acids was observed under SM 25 treatment. Data 620
are presented as mean ± standard error mean (n = 4). Significant differences between groups 621
were determined by statistical analysis: ns = not significant, *p < 0.05, **p < 0.01, ***p < 622
0.001, ****p < 0.0001. 623
624
625
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

626
Figure 4: Kinetic drought exposure modulates sugar metabolism in potato leaves . 627
Quantitative analysis of selected metabolites derived from GC -MS data across different soil 628
moisture conditions: Control, SM 50, SM 25, and SM 0. The levels of soluble sugars (fructose, 629
glucose, erythrose, sucrose, Mannitol , glucopyranoside) are altered in the mild stress (SM 50). 630
Data are presented as mean ± standard error mean (n = 4). Significant differences between 631
groups were determined by statistical analysis: ns = not significant, *p < 0.05, **p < 0.01, ***p 632
< 0.001, ****p < 0.0001.633
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

634
Figure 5: Drought induced metabolic variations in potato leaves. Relative abundance of metabolic change drought treatment was normalized 635
with internal standard and average were means of four biological replicates. The altered pathways are shown as black colored boxes. The bar 636
represents the Control (CC) and drought treatments (SM 50,25 and 0) metabolic response. The response of metabolites in the pathway is represented 637
from Min to Max; dark red color has more abundance while green color is mapped to lower abundance. Altogether, few metabolite s have similar 638
responses to drought stress while levels of metabolites in GS/GOGAT, sugar and fatty acid metabolism were significantly changed.639
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint

640
641
Figure 6: Drought biomarkers towards smart irrigation advisories, GC-MS, colorimetric, and biochemical based detection of drought stress 642
for irrigation advisory. A) GC-MS spectra showing the relative abundance of metabolites Proline, Valine, Serine, and Aspartate under different soil 643
moisture regimes, highlighting proline accumulation under drought stress. B) Colorimetric difference among the samples C) Proline content (µg/ml) 644
in samples. Bars represent the mean values, with different symbols (circle, square, triangle, pentagon) indicating statistica lly significant differences 645
between treatments (± standard error mean) **** =highly significant, ns = nonsignificant. D)A schematic representation of translation of assay 646
adaptable for water stress diagnostics in future, aimed at supporting farmer-oriented irrigation advisory. 647
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.19.712810doi: bioRxiv preprint