Cover crops can provide many benefits to growers, like improving water infiltration and reducing nutrient loss. However, growers in California’s southern San Joaquin Valley worry that the lack of consistent winter rainfall and high cost of water in the area make cover cropping impractical (Mitchell et al. 2017).

In this article, we’ll discuss how well different cover crop mixes suppressed weeds, provided resources for beneficial insects and improved water infiltration. We’ll also delve into water requirements and effects on soil nitrate.

Location of Trials & Cover Crop Species Selection

These trials took place at UC-managed research farms in Shafter (Kern County) and Parlier (Fresno County). The Parlier research farm is the Kearney Agricultural Research and Extension Center and will be referred to as “Kearney” for the remainder of this article.

Four common cover crop mixes were planted at Kearney and five mixes were planted at Shafter (Table 1). Three of the mixes were similar in both locations. We worked with Kamprath Seeds to select cover crop species based on mixes that had done well in a demonstration trial in Shafter the previous year. Two beginner-friendly simple mixes were also evaluated.

Based on the results of our trial, we have outlined some suggestions for trying out cover crops on your farm. Our suggestions depend on your goals for planting cover crops and your concerns about fitting them into your existing cropping system.

Goal: Weed Suppression

Based on the trial results, if your goal is weed suppression, you may consider the following.

Grasses appeared to be most effective at suppressing weeds, especially Merced rye which grew vigorously in both irrigated and non-irrigated plots at both locations. Brassicas also contributed to weed suppression, while legumes were least competitive with weeds, especially at Kearney.

If your goal is weed suppression, consider higher percentages of grasses and brassicas in your seed mix.

Higher seeding rates led to greater weed suppression. If weed suppression is your goal, consider seeding at rates higher than recommended.

If it is not realistic to purchase enough seed for higher seeding rates, pre-irrigating before the first rain to allow weeds to germinate, followed by cultivation prior to planting, can help knock back weeds that may compete with cover crops at germination.

If pre-irrigation is not possible, consider waiting for weed seeds to germinate after the first decent rain, then cultivating your field before planting your cover crop.

However, try to not wait too long, as colder temperatures may inhibit germination. This may have been another reason why cover crop establishment was not as vigorous at Kearney; we planted on December 18 when the high temperature was 53 degrees F and the low was 40 degrees F. Compare this to the planting date of November 5 in Shafter where the high was 20 degrees F warmer (high of 73 degrees F, low of 52 degrees F).

With legume species, in particular, less initial germination correlated with a low percentage of legume species in the final stand. Grasses, however, germinated well in both locations and dominated the final stand.

Therefore, if you cannot plant until later in the winter, consider seeding grass seeds in higher proportions, especially if weed suppression is your goal.

Irrigating directly after seeding may help cover crops establish more vigorously and compete with weeds. If you are not able to irrigate your cover crop, planting right before a rain event can help your cover crops establish a better stand.

Do not assume that there is enough moisture in the soil to help your cover crops properly germinate if you are planting after a rain event.

Goal: Provide Resources for Beneficial Insects

The legumes in both locations did not perform well, so they did not offer much sustenance for pollinators.

The brassicas in Shafter started blooming at the beginning of February, while the radishes at Kearney started blooming in mid-March. This provided forage and diversity for the pollinating bees from the surrounding almond orchards. Butterflies were also observed on these flowers.

Ladybugs were seen on the grasses in both locations, which can serve to keep aphid populations low on surrounding or subsequent crops.

Goal: Improve Water Infiltration

At Kearney, water infiltration was measured using a mini disk infiltrometer before planting cover crops and right before termination of cover crops.

Infiltration rates were higher when cover crops were in the ground compared to the bare soil prior to planting cover crops (Figure 1).

There were no significant differences in infiltration across seed mixes. Therefore, based on our results, there is not one particular mix we would recommend to improve water infiltration.

Having roots in the ground improves the ability of water to enter the soil surface. Roots create channels for water to enter so that more water can be stored in the soil.

With lower infiltration rates, water is more likely to run off the surface. This is why avoiding bare soil is important if your goal is to increase the amount of water that can be stored in your soil.

Concern: Will the Cover Crops Tie Up Soil Nitrogen?

In Shafter, we took soil samples zero to six inches deep on April 21 in the irrigated plots five weeks after termination. The fallow samples had an average soil nitrate level of 45.8 mg/kg, which was similar to the soil nitrate level that we measured before planting the cover crops.

Two of the mixes had soil nitrate levels that were very close to that of the fallow area: the brassica pollinator mix and the soil builder mix. Those mixes were mostly or entirely brassicas, so they decomposed quickly enough that the nitrate they had taken up had returned to the soil. The other three mixes (soil health, rye and peas, and barley and vetch) led to significantly lower soil nitrate levels.

Grasses have higher C:N ratios, which meant that soil microbes took longer to decompose their residue and needed to mine soil nitrogen to do so.

Concern: Can I Really Grow Cover Crops Without Irrigation?

Shafter
All of the non-irrigated plots contributed some amount of biomass. This means that even if you cannot irrigate your cover crops, you can still reap some soil health benefits.

For four of the mixes, the irrigated plots produced at least twice as much biomass as the non-irrigated plots. In contrast, the non-irrigated brassica plots produced almost as much biomass as the irrigated brassica plots (Figure 2).

Kearney
The differences in biomass between the irrigated and non-irrigated plots were not significant, except for the irrigated Merced rye and peas and the irrigated barley and vetch. These two irrigated mixes yielded nearly twice as much biomass as their non-irrigated counterparts (Figure 3). This biomass was mostly attributable to the grasses in both mixes.

The non-irrigated side produced an impressive amount of biomass given that it only received 3.71 inches of water.

The mixes that performed best without irrigation were the Merced rye and peas and the soil builder mix. The daikon radish performed well in the soil builder mix without supplemental irrigation, and the Merced rye performed well in the rye and peas mix without supplemental irrigation. Therefore, Merced rye and radish would be good options if you are unable to irrigate your cover crops.

Conclusion

Based on the results of our trial, cover crops can still produce biomass without irrigation in the San Joaquin Valley, even during a drier winter. While they may not produce huge amounts of biomass without irrigation, having roots in the ground is still beneficial for suppressing weeds, providing beneficial insect habitat, increasing water infiltration and feeding your soil biology.

Cover crops turn sunlight into carbon that feeds your soil microbes through their root system. Therefore, by simply maintaining living ground cover, you are feeding your soil. In addition, when you return cover crop residues to your soil, microbes will decompose them and slowly release available nutrients that can be used by your cash crop.

If you are growing cover crops in a limited water environment, consider planting Merced rye and brassica species, such as daikon radish or mustards. These are also good options if you plant your cover crops later in the winter.

Triticale also grows well with limited water and decomposes more quickly than Merced rye, which is something to consider if you are disking in your cover crop and planting an annual crop shortly after.

In both Shafter and Kearney, the legume species did not perform well because of the high soil nitrate levels and low water conditions. If your field is in a similar situation, it might be better to save money and not buy legume seeds.

You should also think about the cash crop that will follow the cover crop. For example, you do not want to plant a brassica cash crop after a brassica cover crop. This might increase pest pressure for your brassica cash crop.

Interested in trying out cover crops? The USDA and the California Department of Food and Agriculture have programs to help you pay for it. For more information about the USDA resources, reach out to your local USDA Natural Resource Conservation Service office. For more information about the CDFA Healthy Soils Program or getting started with cover crops, reach out to Shulamit Shroder at sashroder@ucanr.edu or Jessie Kanter at jakanter@ucanr.edu.

Resources

Mitchell, J. P., Shrestha, A., Mathesius, K., Scow, K. M., Southard, R. J., Haney, R. L., Schmidt, R., Munk, D.S. & Horwath, W. R. (2017). Cover cropping and no-tillage improve soil health in an arid irrigated cropping system in California’s San Joaquin Valley, USA. Soil and Tillage Research, 165, 325-335.

Citrus Huanglongbing (HLB, citrus greening disease) is caused by the vector-transmitted phloem-limited bacterium Candidatus Liberibacter asiaticus (CLas). It is the most destructive disease which infects all commercial citrus varieties and threatens citrus industries worldwide (Bove 2006; Graham et al. 2020). Current management strategies include insecticide application to control the transmission vector Asian citrus psyllids (ACP) and antibiotics treatment to inhibit CLas (Barnett et al. 2019), but neither could control HLB effectively. Since the first report of HLB in Florida in 2005, citrus acreage and production in Florida decreased by 38% and 74%, respectively (Graham et al. 2020; Stokstad 2012). The disease has spread to citrus-producing states, including Texas and California. In severely affected areas, such as Florida, effective therapy is demanded because disease eradication is impractical. In recently impacted areas, such as California, preventing new infections is most urgent. Hence, innovative therapeutic and preventive strategies to combat HLB are urgently needed to ensure the survival of the citrus industry.

One of the most effective and eco-friendly strategies for disease management is to utilize plant innate immunity-related genes from disease-resistant or tolerant varieties for plant protection. Upon pathogen infection, plant defense response genes undergo expression reprogramming to trigger plant innate immunity. Plant endogenous small RNAs play a pivotal role in this regulatory process (Huang et al. 2019; Zhao et al. 2013b), including phytohormone- and chemical-induced systemic acquired resistance or defense priming, which can promote robust host immune responses upon subsequent pathogen challenges (Brigitte et al. 2017; Zheng Qing and Xinnian 2013).

Our Approaches

HLB tolerance has been observed in some hybrids (e.g., US-942 and Sydney hybrid 72 (Albrecht and Bowman 2011, 2012a)) or citrus relatives (e.g., Microcitrus australiasica, Eremocitrus glauca and Poncirus trifoliata) (Ramadugu et al. 2016). By comparative analysis of small RNA profiles and the target gene expression between HLB-sensitive cultivars and HLB-tolerant citrus hybrids and relatives (Albrecht and Bowman 2011, 2012a), we identified a list of candidate natural defense genes potentially responsible for HLB tolerance (Huang et al. 2020). One of the candidate regulators is a novel anti-microbial peptide (AMP), which we named “stable antimicrobial peptide” (SAMP). Here, we demonstrate that SAMP not only has antimicrobial activity but also has priming activity and can induce citrus systemic defense responses. This dual-functional SAMP can reduce CLas titer, suppress disease symptoms in HLB-positive trees and activate plant systemic defense responses against new infection.

Results

Through the comparative expression analysis of small RNAs and the target genes between HLB-sensitive cultivars and HLB-tolerant citrus US-942 (Poncirus trifoliata x Citrus reticulata) and microcitrus Sydney hybrid 72 (Microcitrus virgate from M. australis × M. australasica) (Huang et al. 2020), we identified a list of candidate plant immune regulators that are potentially contributable to HLB-tolerance. One candidate regulator is a 67-amino acid (aa) peptide, SAMP, that was predicted with antimicrobial activity (Park et al. 2007). SAMP has significantly higher expression levels in both HLB-tolerant hybrids US-942 and Syd 72 than the HLB-susceptible control trees. We further cloned SAMP genes from HLB-tolerant citrus relatives. We found that SAMP transcripts are closely related and have a significantly higher expression level in HLB-tolerant varieties. We further detected the 6.7kD SAMP in the phloem-rich tissue; bark peels of HLB-tolerant Ma and Pt but not in the susceptible Cs. These results support that the SAMPs are likely associated with the HLB-tolerance trait. According to our functional analysis of SAMP, we list the advantages of using SAMP to manage citrus HLB as the following:

SAMP has bactericidal activity and is heat stable.

We screened SAMPs from several citrus relatives using a C. Liberibacter solanacearum (CLso)/potato psyllid/Nicotiana benthamiana interaction system to mimic the natural transmission and infection circuit of the HLB complex. We found that the SAMP from Ma Australian finger lime (MaSAMP) had the strongest effect on suppressing CLso disease and inhibiting bacterial growth in plants. To directly determine the bactericidal activity of MaSAMP on Liberbacter spp, we developed a viability/cytotoxicity assay of Lcr, a close culturable relative of the CLas and CLso (Fagen et al. 2014; Leonard et al. 2012; Merfa et al. 2019). Using this assay, we found that MaSAMP can rapidly kill the bacterial cells within five hours, which is more efficient than the bactericidal antibiotic, Streptomycin. While the heat sensitivity of antibiotics is a major drawback for controlling CLas in citrus fields, we found that SAMPs are surprisingly heat stable. A prolonged exposure to extreme temperatures of 60 degrees C for 20 hours had minimal effect on MaSAMP, which still retained most of its bactericidal activity, whereas Streptomycin lost its antibacterial activity following the same temperature incubation. Thus, SAMP is a heat-stable, plant-derived antimicrobial peptide that can directly kill Lcr and suppress CLso in plants.

SAMP suppresses CLas in HLB-positive citrus trees.

To determine whether MaSAMP can also suppress CLas in citrus trees, we used the pneumatic trunk injection method to deliver the MaSAMP solution into the HLB-positive trees. We first tested with eight CLas-positive Citrus macrophylla with similar bacterial titer and disease symptoms for the treatment. At eight weeks following two doses injected, separated by two months of MaSAMP injection, the disease symptoms and the bacterial titer in all six treated trees were drastically reduced compared to the mock-treated (buffer only) plants and one tree with no CLas detected. We further tested with HLB-positive ‘Madam Vinous’ sweet oranges and Lisbon Lemon trees; those appeared to have declining symptoms and similar CLas titer. After MaSAMP treatments, the trees had developed symptomless new flushes, while mock trees exhibited symptomatic flushes (Fig. 1). The CLas titer was reduced in MaSAMP-treated trees, while it increased in the mock-treated trees. Taken together, these results demonstrate across three trials that SAMP injection can suppress CLas titer in three different HLB-susceptible citrus varieties and can cause trees in declining health to recover.

SAMP treatment safeguards healthy citrus trees from CLas infection.

Protecting healthy trees from CLas infection is critical for managing HLB. The establishment of defense priming in plants can promote faster and/or stronger host immune responses upon pathogen challenges (Brigitte et al. 2017; Zheng Qing and Xinnian 2013). To determine whether MaSAMP has priming activity, we applied it by foliar spray to citrus plants. We found that MaSAMP applications triggered prolonged induction of defense response genes. Thus, SAMP can potentially “vaccinate” uninfected citrus trees and induce defense responses to combat against HLB. To test the protection ability of SAMP on citrus trees, we applied the MaSAMP solution or buffer as mock treatment by foliar spray onto young, healthy ‘Madam Vinous’ sweet orange trees. Five days after treatment, the trees were exposed to ACP carrying CLas under the “no choice feeding” condition for 21 days. We treated trees with MaSAMP solution by foliar spray every two months subsequently. The result indicates that MaSAMP-treated trees have a lower infection rate.

SAMP disrupts the outer membrane and causes cell lysis of the bacterial cell.

To understand the mechanism of MaSAMP bactericidal activity, morphological changes of Lcr post-MaSAMP treatment were observed using transmission electron microscopy. Application of 10 μM MaSAMP to Lcr caused cytosol leakage and the release of small extracellular vesicles after 30 minutes of incubation. The Lcr cells were lysed within two hours of incubation. We isolated the membrane fraction from the MaSAMP treated Lcr and detected the enrichment of MaSAMP in the outer membrane fraction compared with the inner membrane fraction. Thus, MaSAMP likely disrupts mainly the outer membrane of Lcr and breaks the bacterial cells, which leads to cell lysis.

SAMP has Low toxicity.

Because SAMP is internalized by citrus, it is important to test its phytotoxicity. We injected different concentrations of MaSAMP solution directly into citrus leaves and found that MaSAMP has little phytotoxicity. Furthermore, we found that MaSAMP can be detected in fruit tissue of both Australian finger lime and trifoliate orange by Western blot analysis and is very sensitive to human endopeptidase Pepsin, a major gastric enzyme produced by stomach chief cells. Thus, MaSAMP in Australian finger lime has already been consumed by humans for hundreds of years and can be easily digested (Figure 2). These results suggest a low possibility of toxicity of SAMP on citrus and humans, although additional safety assessment tests are necessary for regulatory approval.

Conclusion

Current methods for HLB management include insecticidal control of the vector (Stansly et al. 2014), antibacterial treatments (Blaustein et al. 2018; Gottwald 2010; Hu et al. 2018; Zhang et al. 2014) and nutrient supplements (Rouse 2013; Zhao et al. 2013a). The overuse of insecticides and antibiotics is known to pose threats to human and animal health and select for resistance in the target insect population (Tiwari et al. 2011). Further, current bactericidal or bacteriostatic treatments mostly involve sprays of antibiotics, such as streptomycin and oxytetracycline, which are likely to select for antibiotic-resistant bacteria strains and disrupt the citrus microbiome and ecosystem and may further affect the effectiveness of these antibiotics for medical antibacterial treatment in humans and animals.

On the contrary, SAMPs have a distinct mode of action and tend to interact with the bacterial cell membrane through nonspecific mechanisms, making the emergence of resistant bacteria less likely (Jochumsen et al. 2016; Rodriguez-Rojas et al. 2014). Moreover, SAMP kills bacteria faster than antibiotics, reducing bacterial generations and further lowering the possibility of evolved resistance (Fantner et al. 2010). Most importantly, the heat stability of SAMP can provide a prolonged and durable effect in the field compared to heat-sensitive antibiotics. SAMP not only kills bacteria cells but can also prime plant immune responses to prevent/reduce infection. In our greenhouse trials, SAMP has been shown to treat HLB-positive trees and inhibit the emergence of new HLB-infection in healthy trees. Field trials, which can take several years, are currently being initiated in Florida to confirm the efficacy of SAMP in controlling HLB. Field trials also include testing multiple peptide application methods for citrus growers to prevent and treat HLB.

Contact Hailing Jin at hailingj@ucr.edu for more information.

References:

Albrecht, U., and Bowman, K.D. (2011). Tolerance of the trifoliate citrus hybrid US-897 (Citrus reticulata x Poncirus trifoliata) to huanglongbing. HortScience 46, 16-22.

Albrecht, U., and Bowman, K.D. (2012a). Tolerance of trifoliate citrus hybrids to Candidatus liberibacter asiaticus. . Sc Horticulturae 147, 71-80.

Barnett, M.J., Solow-Cordero, D.E., and Long, S.R. (2019). A high-throughput system to identify inhibitors of Candidatus Liberibacter asiaticus transcription regulators. Proceedings of the National Academy of Sciences of the United States of America 116, 18009-18014.

Blaustein, R.A., Lorca, G.L., and Teplitski, M. (2018). Challenges for Managing Candidatus Liberibacter spp. (Huanglongbing Disease Pathogen): Current Control Measures and Future Directions. Phytopathology 108, 424-435.

Bove, J.M. (2006). Huanglongbing: a destructive, newly-emerging, century-old disease of citrus. . J Plant Pathol 88, 7-37.

Brigitte, M.-M., Ivan, B., Estrella, L., and Victor, F. (2017). Defense Priming: An Adaptive Part of Induced Resistance. Annual review of plant biology 68, 485-512.

Fagen, J.R., Leonard, M.T., Coyle, J.F., McCullough, C.M., Davis-Richardson, A.G., Davis, M.J., and Triplett, E.W. (2014). Liberibacter crescens gen. nov., sp. nov., the first cultured member of the genus Liberibacter. International journal of systematic and evolutionary microbiology 64, 2461-2466.

Fantner, G.E., Barbero, R.J., Gray, D.S., and Belcher, A.M. (2010). Kinetics of antimicrobial peptide activity measured on individual bacterial cells using high-speed atomic force microscopy. Nature nanotechnology 5, 280-285.

Gottwald, T.R. (2010). Current epidemiological understanding of citrus Huanglongbing. Annu Rev Phytopathol 48, 119-139.

Graham, J., Gottwald, T., and Setamou, M. (2020). Status of Huanglongbing (HLB) outbreaks in Florida, California and Texas. Trop Plant Pathol.

Hu, J., Jiang, J., and Wang, N. (2018). Control of Citrus Huanglongbing via Trunk Injection of Plant Defense Activators and Antibiotics. Phytopathology 108, 186-195.

Huang, C., Niu, D., Kund, G., Jones, M., Albrecht, U., Nguyen, L., Bui, C., Ramadugu, C., Bowman, K., Trumble, J., et al. (2020). Identification of citrus defense regulators against citrus Huanglongbing disease and establishment of an innovative rapid functional screening system. Plant Biotechnology Journal.

Huang, C.Y., Wang, H., Hu, P., Hamby, R., and Jin, H. (2019). Small RNAs – Big Players in Plant-Microbe Interactions. Cell host & microbe 26, 173-182.

Jochumsen, N., Marvig, R.L., Damkiaer, S., Jensen, R.L., Paulander, W., Molin, S., Jelsbak, L., and Folkesson, A. (2016). The evolution of antimicrobial peptide resistance in Pseudomonas aeruginosa is shaped by strong epistatic interactions. Nature communications 7, 13002.

Leonard, M.T., Fagen, J.R., Davis-Richardson, A.G., Davis, M.J., and Triplett, E.W. (2012). Complete genome sequence of Liberibacter crescens BT-1. Standards in genomic sciences 7, 271-283.

Merfa, M.V., Perez-Lopez, E., Naranjo, E., Jain, M., Gabriel, D.W., and De La Fuente, L. (2019). Progress and Obstacles in Culturing ‘Candidatus Liberibacter asiaticus’, the Bacterium Associated with Huanglongbing. Phytopathology 109, 1092-1101.

Park, S.C., Lee, J.R., Shin, S.O., Park, Y., Lee, S.Y., and Hahm, K.S. (2007). Characterization of a heat-stable protein with antimicrobial activity from Arabidopsis thaliana. Biochemical and biophysical research communications 362, 562-567.

Ramadugu, C., Keremane, M.L., Halbert, S.E., Duan, Y.P., Roose, M.L., Stover, E., and Lee, R.F. (2016). Long-Term Field Evaluation Reveals Huanglongbing Resistance in Citrus Relatives. Plant Dis 100, 1858-1869.

Rodriguez-Rojas, A., Makarova, O., and Rolff, J. (2014). Antimicrobials, stress and mutagenesis. PLoS pathogens 10, e1004445.

Rouse, B.B. (2013). Rehabilitation of HLB Infected Citrus Trees using Severe Pruning and Nutritional Sprays. Proc Fla State Hort Soc 126, 51-54.

Stansly, P.A., Arevalo, H.A., Qureshi, J.A., Jones, M.M., Hendricks, K., Roberts, P.D., and Roka, F.M. (2014). Vector control and foliar nutrition to maintain economic sustainability of bearing citrus in Florida groves affected by huanglongbing. Pest Manag Sci 70, 415-426.

Stokstad, E. (2012). Agriculture. Dread citrus disease turns up in California, Texas. Science 336, 283-284.
Tiwari, S., Mann, R.S., Rogers, M.E., and Stelinski, L.L. (2011). Insecticide resistance in field populations of Asian citrus psyllid in Florida. Pest management science 67, 1258-1268.

Zhang, M., Guo, Y., Powell, C.A., Doud, M.S., Yang, C., and Duan, Y. (2014). Effective antibiotics against ‘Candidatus Liberibacter asiaticus’ in HLB-affected citrus plants identified via the graft-based evaluation. PloS one 9, e111032.

Zhao, H., Sun, R., Albrecht, U., Padmanabhan, C., Wang, A., Coffey, M.D., Girke, T., Wang, Z., Close, T.J., Roose, M., et al. (2013a). Small RNA profiling reveals phosphorus deficiency as a contributing factor in symptom expression for citrus huanglongbing disease. Mol Plant 6, 301-310.

Zhao, H., Sun, R., Albrecht, U., Padmanabhan, C., Wang, A., Coffey, M.D., Girke, T., Wang, Z., Close, T.J., Roose, M., et al. (2013b). Small RNA Profiling Reveals Phosphorus Deficiency as a Contributing Factor in Symptom Expression for Citrus Huanglongbing Disease. Molecular Plant 6, 301-310.

Zheng Qing, F., and Xinnian, D. (2013). Systemic Acquired Resistance: Turning Local Infection into Global Defense. Annual review of plant biology 64, 839-863.

The global population is projected by the United Nations to reach 9.7 billion people by 2050. To satisfy the needs for food and fiber for that many people, agricultural production should increase by >70%. Globally and here in the U.S., studies suggest that improved management of soil resources, irrigation water and nutrients through digital agriculture (See Sidebar Digital Agriculture on page 9.) tools can be the key to increasing agricultural production (Fig. 1). According to USDA, sustaining the economic viability of agriculture is only one of the goals that must be achieved to fully satisfy the long-term food and fiber needs of the American people. Other key goals are to reduce the environmental footprint of agriculture and to improve the quality of life for farm families and communities.

Improving Production Systems

Agriculture in the Southwestern U.S. is vital to the prosperity of rural communities and the national economy, both for internal consumption and for international export. Many crops are grown in the Southwestern U.S., including field crops and many specialty fruits, vegetables and nuts. In particular, around $12 billion/year in income is generated by agriculture in California’s Salinas River Valley and in the Colorado River Basin, including regions in Southern California that use Colorado River water for irrigation. In these regions, agriculture employs more than 500,000 workers.

Sustaining and improving the agricultural productivity in the region in the long term is, however, under threat as growers face major challenges: climate change, disease and pest outbreaks, increasing salinity and diminished and/or degraded soil and water resources, to mention a few. The region has experienced prolonged and major droughts since 2000, with the possibility of streamflow reductions by more than 50% by 2100. Future requirements to preserve groundwater may remove 500,000 acres from production in California alone. Minimizing salt and nutrient loading is increasingly mandated for effective water reuse. Uncertainty in the water supply in these regions is particularly troublesome, as agriculture there consumes 39% of the U.S. total irrigation water (almost 32 million acre-feet per year.) Additionally, climate change threatens to increase insect, pathogen and weed pressures and geographic distribution, while public interest and increasing on-farm costs push toward reducing pesticide use and organic farming.

Our current understanding of agricultural production systems indicates that Crop Yield is a complex function of Genetics × Environment × Management × Space × Time interactions. Understanding why yield deviates from optimal over space and time in different landscapes is key to adjusting management cost-effectively. When looking at the multi-year productivity of farmland, a rule of thumb might say that yield varies over time in response to climate as much as it changes across different spatial scales (within a field and across multiple fields) due to the variability in soil and landscape features.

Failure to adapt management to the dynamic spatial and temporal variability of crop growth often results in crop loss or over-application of agronomic inputs, which can lead to economic loss and environmental degradation. Real-time crop growth models that can leverage information from very high spatial and temporal resolution satellite imagery and ground networks of sensors, such as weather stations, are great candidates for guiding site-specific tailored agronomic management. Moreover, when combined with artificial intelligence, crop growth models can help improve farm management while considering the tradeoffs between different sustainability aspects at different spatial scales. For example, how to increase field scale profitability while reducing regional-scale environmental impacts.

Current Research

With the overarching goal of increasing agricultural profitability by reducing and optimizing inputs to increase yield and curb losses from abiotic and biotic stressors in the Southwestern U.S., UC Riverside recently started a five-year project on the use of artificial intelligence and big data from high-resolution imagery and ground sensor networks to improve the management of irrigation, fertilization and soil salinity as well as to enable early detection of weeds and pests. The project is led by Elia Scudiero, a professional researcher in UC Riverside’s Department of Environmental Sciences, and includes several co-investigators at UC Riverside and UC ANR, USDA-ARS, University of Arizona, Duke University, Kansas State University and University of Georgia. To accomplish its goal, the project relies on many collaborations with ag tech industry partners, including Planet Labs, Inc. Through the collaboration with Planet Labs, the project investigators will use daily high-resolution (around 12 feet) satellite imagery to monitor crop growth and soil properties.

Some of the artificial intelligence applications that the project will develop, for a variety of crops, include crop inventorying, soil mapping, estimations of crop water use and requirement, estimation of plant and soil nutrient status, analyzing the feasibility and cost-effectiveness of variable-rate fertigation across selected regions in the Southwestern U.S. and detecting weeds and pathogens. In the remainder of this article, we will provide a general overview and some preliminary results from some selected applications: Water use and water requirement estimations, mapping soil salinity with remote sensing and detecting biotic stressors.

Water Use and Water Requirement Estimations

A key element of this project will be to provide reliable estimates of crop water use and irrigation forecasting at a very high resolution daily. Current remote-sensing-based evapotranspiration models often suffer from infrequent satellite overpasses, which are generally available weekly or every two weeks at the 30- to 100-foot spatial resolution. Such sporadic information is a limitation, especially for vegetable crops, which can have very fast-growing cycles. To overcome this and other limitations, the project is integrating daily meteorological information (from state and federal networks) and high-resolution satellite data from Planet Labs with BAITSSS, an evapotranspiration (ET) computer model developed by Dr. Ramesh Dhungel (a Research Scientist in the project) and colleagues. The current version of BAITSSS leverages information from the Landsat 8 satellite platform (NASA). An example application of the model is shown in Figure 1, where BAITSSS is combined with soil moisture modeling to forecast when irrigation is needed to supply crop water demand.

The project’s lead on water use and requirement estimations, Dr. Ray Anderson (USDA-ARS U.S. Salinity Laboratory, Riverside, Calif.), says that “along with irrigation management, we believe the daily imagery from Planet Labs will allow growers and irrigation managers to see evapotranspiration anomalies within a field. These anomalies could indicate irrigation issues or decreases in plant health due to other abiotic stressors such as salinity or nutrient deficiency. Identification of these anomalies will ensure more efficient field scouting and earlier identification of issues before permanent yield loss occurs.”

Mapping Soil Properties

Accurate knowledge of spatial variability of soil properties, such as texture, hydraulic properties and salinity, is important to best understand the reasons of crop yield spatial variability. Scudiero’s Digital Agronomy Lab at UC Riverside is developing novel tools based on machine learning to automate near-ground sensing of soil properties and remote sensing of soil salinity in irrigated farmland. Soil salinity maps are very useful to inform field-scale irrigation practices (e.g., to calculate the amount of irrigation water needed to avoid harmful accumulation of salts in the soil profile.)

In this project, Scudiero and his team will be using field-scale soil maps of soil salinity collected in the past (since the early 1980s) and throughout the next five years by UC Riverside, the USDA-ARS U.S. Salinity Laboratory, the University of Arizona and other collaborators to generate soil salinity maps for the entire Southwestern U.S. Figure 3 shows the soil salinity map produced for the western San Joaquin Valley by Scudiero and colleagues (see the additional resources section.) In particular, Scudiero’s team will use the ground information to calibrate the Planet Labs time-series imagery to predict soil salinity in the root zone (e.g., the top four feet of the soil profile.) Satellite imagery alone is generally not sufficient to predict soil salinity. Other stressors (water stress, nutrient deficiency) have similar imagery properties to salinity. However, in short periods (two to five years), salinity remains fairly stable throughout the soil profile, contrary to other more transient stressors. Because of these differences in temporal variability between stressors, multi-year time series can be used to detect and map crop health reduction due to soil salinity.

Biotic Stress Detection

Within this project, UC Riverside scientists are using smartphone pictures, drone imagery and Planet Labs imagery (20 inch and 12 feet resolution) to develop artificial intelligence classifiers for early biotic stress detection. Figure 4 shows an example of a survey carried out by project collaborators Sonia Rios (UC ANR) and Robert Krueger (USDA-ARS) at UC Riverside’s Coachella Valley Agricultural Research Station in Thermal, Calif. Preliminary analyses show that 20×20-inch resolution imagery from the Planet SkySat satellites can be used to identify the presence of weeds against bare soil ground coverage in citrus and date palm. Single-date imagery cannot be used to successfully distinguish between different weed species. The project investigators hope that repeated satellite imagery, together with drone imagery, will be successful at identifying the emergence and extent of weeds and other biotic stress at the field and farm scales. To develop such tools, project investigators are carrying out controlled experiments at the UC Riverside research farms with controlled weed and pathogen pressures on a variety of vegetable crops.

Training to Growers and Consultants

From conversations with stakeholders in California and other states in the Southwestern U.S., we learned that most growers are committed to maintaining the quality and profitability of soil, water and other natural resources in the long term. Additionally, many growers recognize the potential of integrating digital agriculture technologies in their daily decision-making. Nevertheless, investing in new technology always requires considerable commitment. Therefore, growers would like to have more information on the cost-effectiveness of state-of-the-art technologies and on the suitability of technology to their local agricultural system. To address these and other questions, the project team is establishing a multi-state cooperative extension network to develop training programs for growers and consultants on the topics of precision agriculture, digital agriculture and the use of soil and plant sensors in agriculture. These training activities, scheduled to start in 2021, will include contributions from university personnel and industry members.

This article provided an overview and some preliminary results for the University of California Riverside, or UCR, -led project on “Artificial Intelligence for Sustainable Water, Nutrient, Salinity, and Pest Management in the Western US”. The project is funded by U.S. Department of Agriculture’s National Institute of Food and Agriculture (Grant Number: 2020-69012-31914). Through September 2025, the project will investigate the use of daily high-resolution satellite imagery and data science to identify inefficiencies in agronomic management practices and to support improved irrigation, fertilization and pest control in the irrigated farmland across the Colorado River Basin and Central and Southern California.

Additional information on the project and the content of this article can be requested from Elia Scudiero (elia.scudiero@ucr.edu). Information about the project’s cooperative extension events can be requested from ai4sa@ucr.edu. Further information about research on remote sensing of soil salinity and evapotranspiration can be found in the additional resources section.

Additional Resources:
Dhungel, R., R.G. Allen, R. Trezza, and C.W. Robison. 2016. Evapotranspiration between satellite overpasses: methodology and case study in agricultural dominant semi-arid areas: Time integration of evapotranspiration. Meteorol. Appl. 23(4): 714–730. doi: 10.1002/met.1596.
Scudiero, E., Corwin, D.L., Anderson, R.G., Yemoto, K., Clary, W., Wang, Z.L., Skaggs, T.H., 2017. Remote sensing is a viable tool for mapping soil salinity in agricultural lands. California Agriculture 71, 231-238. doi: 10.3733/ca.2017a0009