The model’s applicability was demonstrated with a simple case study

The 2C model does not capture other observed dynamics, such as the down regulation of the Agr system and its role in biofilm formation.These hypotheses predict very different carrier and response probabilities. This is because the former assumes that the adapted state is unaffected by resident microflora while the latter assumes that the adapted and un-adapted states are equally affected. The truth is likely in the middle, i.e., SA in the adapted state are affected by resident microflora to a lesser extent than SA in the unadapted state. Such an approach was not pursued in the spirit of avoiding over-fitting given the limited data. We note that more data collected within the first day of inoculation will help judge the quality of the hypotheses presented in this study, which need to be evaluated on an absolute scale with a goodness of fit test. Confidence in the predictions of the model will improve as more data is gathered, either supporting or refuting the hypotheses. The mechanistic nature of the model enabled direct simulation of repeated exposures from the environment, without having to assume independence between exposure events. This paves the way for more involved modeling efforts such as accounting for healthcare workers and other hospital surfaces that contain MRSA. Such efforts can be challenging for two reasons: the availability of high quality data to model behaviors and the computational effort in simulating stochastic systems. However,aeroponic tower garden system they can supplement our understanding of the environment as a source of MRSA and help devise the most effective control measures in hospitals and the community.

New root phenotyping technological developments are needed to overcome the limitations of traditional destructive root investigation methods, such as soil coring or “shovelomics” . Mancuso and Atkinson et al. provide extensive reviews on the methodological advances on non-destructive root phenotyping, including Bioelectrical Impedance Analysis , planar optodes, geophysical methods, and vibrating probe techniques. These techniques aim to mitigate key limitations of traditional root phenotyping, especially addressing the need for a better and more convenient characterization of the finer roots and root functioning. Advances in non-invasive and in-situ approaches for monitoring of root growth and function over time are needed to gain insight into the mechanisms underlying root development and response to environmental stressors. Geophysical methods have been tested to non-destructively image roots in the field. Ground Penetrating Radar approaches have been used to detect coarse roots . Electrical Resistivity Tomography and Electro Magnetic Induction approaches have been used to image and monitor soil resistivity changes associated with the Root Water Uptake . Recent studies explored the use of multi-frequency Electrical Impedance Tomography to take advantage of the root polarizable nature . Despite these advantages, geophysical methods to date share common limitations regarding root characterization. Geophysical methods developed to investigate geological media: in the case of roots they measure the root response as part of the soil response, see Fig. 1a for the ERT acquisition. Because of the natural soil heterogeneity and variability the resolution and signal characteristics of geophysical methods strongly depend on soil type and conditions. As such, interpretation of the root soil system response is non-unique, hindering the differentiation between roots of close plants and the extraction of specific information about root physiology from the electrical signals. Unlike geophysical methods, the BIA for root investigation developed to specifically target the impedance of plant tissues, limiting the influence of the growing medium.

A practical consequence is that BIA involves the application of sensors into the plant to enhance the method sensitivity. BIA measures the electrical impedance response of roots at a single frequency or over a range of frequencies . The measured BIA responses have been used to estimate root characteristics, such as root absorbing area and root mass . Estimation of these root traits is based on assumptions on the electrical properties of roots . A key assumption is that current travels and distributes throughout the root system before exiting to the soil , with no leakage of current into the soil in the proximal root position . It is only in the former case, that the BIA signal would be sensitive to root physiology. Despite the physiological relevance of the BIA assumptions and the number of BIA studies, a suitable solution for the characterization of the current pathways in roots is missing. Thus far, only indirect information obtained from invasive and time-consuming experiments have been available to address this issue . Mary et al. , and Mary et al. 2020 tested the combined use of ERT and Mise A La Masse methods for imaging grapevine and citrus roots in the field. An approach hereafter called inversion of Current Source Density was used to invert the acquired data. The objective of this inversion approach is to image the density and position of current passing from the plant to the soil. The current source introduced via the stem distributes into “excited” roots that act as a distributed network of current sources . Consequently, a spatial numerical inversion of these distributed electric sources provides direct information about the root current pathways and the position of the roots involved in the uptake of water and solutes. The numerical approach used to invert for the current source density is a key component required for such an approach. Mary et al. used a nonlinear minimization algorithm for the inversion of the current source density.

The algorithm consisted of gradient-based sequential quadratic programming iterative minimization of the objective functions described in Mary et al. . The algorithm was implemented in MATLAB, R2016b, using the fmincon method. Because no information about the investigated roots was available, the authors based these inversion assumptions and the interpretation of their results on the available literature data on grapevine root architecture. Consequently, Mary et al. highlighted the need for further iCSD advances and more controlled studies on the actual relationships between current flow and root architecture. In this study, we present the methodological formulation and evaluation of the iCSD method,dutch buckets for sale and discussits applications for in-situ characterization of current pathways in roots. We perform our studies using laboratory rhizotron experiments on crop roots. The main goals of this study were: 1) develop and test an iCSD inversion code that does not rely on prior assumptions on root architecture and function; 2) design and conduct rhizotron experiments that enable an optimal combination of root visualization and iCSD investigation of the current pathways in roots to provide direct insight on the root electrical behavior and validate the iCSD approach; and 3) perform experiments to evaluate the application of the iCSD method on different plant species and growing medium that are common to BIA and other plant studies.The relationship between hydraulic and electrical pathways has been the object of scientific debate because of its physiological relevance and methodological implications for BIA methods . A key and open question concerns the distribution of the current leakage . The distribution of the current leakage is controlled by 1) the electrical radial and longitudinal conductivities , and 2) by the resistivity contrast between root and soil. With regard to σcr and σcl, when σcl is significantly higher than σcr, the current will predominantly travel through the xylems to the distal “active” roots, which are mostly root hairs. Based on the link between hydraulic and electrical pathways, this is consistent with a root water uptake process where root hairs play a dominant role while the more insulated and suberized roots primarily function as conduits for both water and electric current . On the contrary, if the σcr is similar to σcl, the electrical current does not tend to travel through the entire root system but rather starts leaking into the surrounding medium from root proximal portions. The coexistence of proximal and distal current leakage is in line with studies that suggest the presence of a more diffused zone of RWU, and a more complex and partial insulation effect of the suberization, possibly resulting from the contribution of the cell-to-cell pathways .

Soil resistivity can affect the distribution of the current leakage by influencing the minimum resistance pathways, i.e., whether roots or soil provide the minimum resistance to the current flow. In addition, soil resistivity strongly relates to the soil water content, which, as discussed, affects the root physiology. Therefore, information on the soil resistivity, such as the ERT resistivity imaging, has the potential for supporting the interpretation of both BIA and iCSD results. Dalton proposed a model for the interpretation of the plant root capacitance results in which the current equally distributes over the root system. Because of the elongated root geometry this model is coherent with the hypothesis of a low resistance xylem pathway . Numerous studies have applied Dalton’s model documenting the predicted correlation between root capacitance and mass . In fact, recent studies with wheat, soy, and maize roots continue to support the capacitance method . Despite accumulating studies supporting the capacitance method, hydroponic laboratory results of Dietrich et al. and other studies have begun to uncover potential inconsistencies with Dalton’s assumptions. In their work, Dietrich et al. explored the effect of trimming submerged roots on the BIA response and found negligible variation of the root capacitance. Cao et al. reached similar conclusions regarding the measured electrical root resistance . Urban et al. discussed the BIA hypotheses and found that the current left the roots in their proximal portion in several of their experiments. Conclusions from the latter study are consistent with the assumption that distal roots have a negligible contribution on root capacitance and resistance. Because of the complexity of the hydraulic and electrical pathways, their link has long been the object of scientific research and debate. For recent reviews see Aroca and Mancuso ; for previous detailed discussions on pathways in plant cells and tissues see Fensom , Knipfer and Fricke , and Findlay and Hope ; see Johnson and Maherali et al. in regard to xylem pathways. See Jackson et al. and Hacke and Sperry for water pathways in roots. Thus, above discrepancies in the link between electrical and hydraulic root properties can be, at least to some degree, attributed to differences among plant species investigated and growing conditions. Among herbaceous plants, maize has been commonly used to investigate root electrical properties . For instance, Ginsburg investigated the longitudinal and radial current conductivities of excited root segments and concluded that the maize roots behave as leaking conductors. Similarly, Anderson and Higinbotham found that σcr of maize cortical sleeves was comparable to the stele σcl. Recently, Rao et al. found that maize root conductivity decreases as the root cross-sectional area increases, and that primary roots were more conductive than brace roots. By contrast, BIA studies on woody plants have supported the hypothesis of a radial isolation effect of bark and/or suberized tissues . Plant growing conditions have been shown to affect both water uptake and solute absorption due to induced differences in root maturation and suberization . Redjala et al. observed that the cadmium uptake of maize roots grown in hydroponic conditions was higher than in those grown aeroponically. Tavakkoli et al. demonstrated that the salt tolerance of barley grown in hydroponic conditions differed from that of soil-grown barley. Zimmermann and Steudle documented how the development of Casparian bands significantly reduced the water flow in maize roots grown in mist conditions compared to those grown hydroponically. During their investigation on the effect of hypoxia on maize, Enstone and Peterson reported differences in oxygen flow between plants grown hydroponically and plants grown in vermiculite. The results reported above and in other investigations are conducive to the hypothesis that root current pathways are affected by the growing conditions, as suggested in Urban et al. . For example, the observations by Zimmermann and Steudle and Enstone and Peterson may explain the negligible contributions to the BIA signals from distal roots under hydroponic conditions . At the same time, the more extensive suberization in natural soil and weather conditions could explain the good agreement between the rooting depth reported by Mary et al. based on the iCSD and the available literature data for grapevines in the field.

Crop water requirements either remain constant or increase for some crops

A table of sensor specifications can be found in Appendix 3. The primary source of ground truth observations for the CDL products is the USDA Farm Service Agency Command Land Unit program . The FSA CLU comprises digitized polygon boundaries of “semi-permanent ‘fields’” and is a confidential NASS-internal data set. Auxiliary input data sources include the USGS National Elevation Data set , the Multi-Resolution Land Characteristics Consortium National Land Cover Dataset . Prior to 2006, classification was performed using a maximum likelihood classifier in the NASS-internal Peditor program, an image processing software written in Pascal and FORTRAN . Beginning in 2006, Rulequest Research’s See5.0 software was used to create a decision tree classifier. This is applied to the remotely sensed imagery using the MRLC NLCD Mapping Tool and ERDAS Imagine. Accuracy reports are presented in state-level metadata files each annual CDL survey. For supervised classification, ground-truthed observations are defined as polygons, and are subsequently buffered inward by 30 meters. This was done in part to reconcile differences between the different spatial resolutions of the remotely sensed imagery . Prior to 2016, this method of inward-buffering was used for validation and the construction of accuracy reports. However, this excluded edge pixels from the accuracy reports. This resulted in a somewhat inflated accuracy assessments. Starting in 2017, only “unbuffered” accuracy assessments are reported. In 2016, the CDL metadata included both “buffered” and “unbuffered” accuracy reports . Overall accuracy for California FSA crops tend to range between 80 and 90 percent. In order to determine the proportion of daily crop water requirements that were met by direct rainfall,livestock fodder system this study used precipitation observations from the Parameter-elevation Regressions on Independent Slopes Model climate mapping system.

Specifically, 800-m daily precipitation rasters were upscaled to 30 meters, using bilinear interpolation . The PRISM Climate Group at the Northwest Alliance for Computational; Science and Engineering at Oregon State University maintains daily 800-m and 4-km raster datasets of precipitation across the 48 conterminous states, spanning back to 1981. The group also maintains raster datasets of temperature , dewpoint temperature, vapor pressure deficit , and 30-year annual “normals” . PRISM rasters are freely available on the PRISM climate group homepage, with the exception of 800-m monthly and daily data, which must be ordered. PRISM precipitation rasters were commissioned by USDA through the Natural Resources Conservation Service to serve as the official spatial climate data sets of the USDA . PRISM rasters are created at a 30-arcsecond spatial resolution and are also available at a 2.5-arcminute resolution , matching previous USDA-NRCS 1961-1990 climate data sets developed in the 1990s. At its core, PRISM is is an interpolation technique that reproduces the spatial climate patterns of the United States, with a particular emphasis on the effect of elevation and slope on precipitation . The method was originally developed by Dr. Christopher Daly of Oregon State University in an attempt to reproduce the process that climatologists used to construct climate maps of the United States . At its core, the model incorporates data from surface weather stations . PRISM utilizes a linear climate-elevation relationship, rather than a multiple regression model due to difficulties in predicting “complex relationships between multiple independent variables and climate”. Instead, weather station observations are weighted by distance, elevation, coastal proximity, topographic facet, vertical layer, topographic position, and effective terrain . Accuracy estimates using single-deletion jackknife cross validation, leave-one-out cross validation, and a 70% prediction interval have been performed on various revisions to the PRISM method . Regional mean absolute error between predicted and observed precipitation and temperature tend to be similar overall and higher in the physiographically complex western United States.

A 2008 evaluation of PRISM for the central California coast saw good agreement between PRISM, WorldClim, and Daymet temperature observations for the central valley of California .The water footprint incorporates the effect of yields on crop water use. Assuming negligible losses of water, the crop water requirement assumed to be equivalent to the actual crop water use. The resultant water footprints can be considered a “best-case scenario”, as inefficiencies in water distribution and application can only increase the actual crop water use, increasing the blue component of the water footprint. Water footprints are expressed in units of cubic meter of water per metric ton of harvested product. From a resource management perspective, the WF of applied water is most valuable for regions that are predominantly reliant on surface water resources. Total WF figures are presented in Appendix 5, sections G-H. From 2008 to 2015, the blue WF was always orders of magnitude higher than the green WF , further demonstrating the minor role of direct rainfall toward satisfying crop water requirements in California . Across large regional extents, the overall water footprint for most hydrologic regions does not vary much year to year, with the exception of isolated fluctuations driven by changes in reported yield . These fluctuations are also visible in the crop specific annual totals . For example, low mint yields in Shasta County in 2013 inflate the 2013 water footprint for mint . This fluctuation is also visible in the regional WF statistic—due to the low intensity of agriculture in the North Lahontan region , the low yield bias on the WF is visible at the regional scale.A 2018 study by the University of California, Davis compared the consumptive use of water by crops in the Sacramento-San Joaquin Delta of California using seven different crop evpotranspiration models. This “Delta ET” study included methods that were based on crop coefficients and methods which are reliant on remotely sensed satellite measurements.

Monthly crop ET values were published along with the region of interest, for the 2015 and 2016 water years. Overall crop ET observations from this study were compared to the monthly mean of the seven ET models from the Delta ET study. Overall, there was general agreement between this study and the methods detailed in the Delta ET study. This study tended to underestimate crop ET each month by no greater than 50% . The highest proportion of underestimation occurred during the winter months. However, due to the small magnitude of wintertime ET,hydroponic nft gully this only resulted in a 22% cumulative underestimation . The Delta ET models include some land cover classes that this study does not model. The SIMS ET model implemented by the Delta ET study does not model semi agricultural/right-of-way and wet herbaceous/sub-irrigated pasture. The results of this study closely match the monthly results from SIMS within 1%.The water footprint can be thought as a measure of the effectiveness of a unit application of water, given yields as the test for effectiveness. Regionally, it is expected that the highest proportion of crop water use would occur in the intensively-cultivated central valley region . Compared to regions less suitable for agriculture . these regions are exceptional in their overall water use. However, they are not exceptional in the water footprint of agricultural activities . For example, Monterey county contained the largest overall average water footprint of agricultural production, in spite of the possessing a small proportion of overall agricultural water requirement. Among agricultural commodities, average water footprints agree with other assessments in terms of rank order of water footprint and overall crop water requirement. For example, nuts and grasses both have a large water footprint and large crop water requirement, compared to other crops modeled in this study . The large proportional crop water requirement could be function of crop-specific ET characteristics, or it could be an artifact of a large overall cultivated area. However, when compared the proportion of harvested acres, lower WF crops make up a slightly larger portion of cultivated acres than fruits and nuts . Droughts can be used to study the effects of reductions the overall amount of water available in a distribution system. For the drought period starting in 2012, reductions in harvested acres were observed, especially with grasses and some specialty crops.The equivalence of crop water use and crop water requirement was a central assumption in this study. The response of the water footprint under deficit irrigation can be an important topic for future work, as reductions in water use may be less effective from a footprint perspective if yields are dramatically affected. This study explored the distribution of water footprints across the State of California regionally, across different commodities, and across a 7 year period, marked by wet and dry climatic extremes. A model of crop water use was coupled with surveyed observations of precipitation, harvest statistics, and a land cover model. Findings from this study revealed an overall insensitivity of the water footprint to climatic extremes and significant inter-annual variability in the metric .

As a highly derivative metric, the water footprint accumulates errors from all of the data sources used in its calculation. Unreliable yield reports can dramatically change the water footprint, due to the power-law relationship between the water footprint and crop yield. By quantifying the uncertainty of this metric, the water footprint could become even more useful as a decision support tool. However, even exploring the relative proportions of water footprints are useful in defining the conceptual extent of the water-use for a given territory or commodity. Future studies can conduct sensitivity analysis of the metric, to examine which input parameters have the greatest effect on water footprint variability. In the course of this study, a framework was created and implemented in R that allows this analysis to be replicated and run with different inputs. This framework can be utilized in future analyses to compare the footprint metric with the ever improving agricultural methodologies found in California, from modeling irrigation efficiencies, to using improved land use surveys and methods of modeling crop evapotranspiration. The framework can also be applied to different regions, provided that there are harvest and crop ET models which adequately characterize the region. An understanding of the water footprint of agricultural production can provide information to the grower who wishes to maximize the economic return of a given volume of water, the state planner who wishes to maximize utility per unit of water allocated, the national administrator who wishes to understand national risks and strengths, or the informed citizen who wishes to align their consumptive activities with a vision for the conditions conferred to the next generation. This information is a critical component of the continuous motivation to characterize relationships between society and natural resource systems, with the ultimate goal of creating sustainable and resilient social and natural systems.Approximately 5 million workers in the United States work in the agricultural industry with potential exposure to a wide variety of respiratory toxicants. Among the various respiratory hazards not well studied is obstructive and restrictive lung disease caused by inorganic minerals . California agricultural workers have increased respiratory symptoms, decreased respiratory function, and increased mortality rates from chronic pulmonary disease compared with the general population . The agricultural environment of the Central Valley of California places individuals at increased risk of exposure to inorganic particles. This region encompasses a rich farming area as well as extensive urban development. The predominantly dry farming techniques of the Central Valley result in high levels of airborne dust from operations such as field preparation and harvesting of row crops and tree fruits . Although soil consists of a mixture of organic and inorganic materials, potential health effects from the ever-present mineral dusts have been largely overlooked. In theory, soil should contain representative portions of all major mineral classes in the earth’s crust; in reality, most agricultural soils are composed largely of silicate materials and crystalline silica  with varying amounts of other mineral classes, depending on the local geologic history. The evidence that mineral dust exposure poses a significant hazard to agricultural workers for interstitial lung disease is based on a handful of case reports, inferences from exposures to mineral dusts in other industries, studies of wild and farm animals exposed to environmental dusts, and toxicologic studies . The deposition and clearance of particles within the respiratory system occurs in an in homogeneous manner.

Higher-value crops including produce and cash crops may also be more sensitive to weather

Few products suitable to agricultural livelihoods are available, and despite the wide proliferation of microfinance institutions , most are limited to non-agricultural activities given there are substantial challenges inherent in long-cycle agricultural lending . Lenders in these contexts charge high interest rates to help offset their assessment of the risk that loans will not be repaid. These higher interest rates can, perversely, have the effect of attracting only borrowers with no intention of repaying , thus driving interest rates even higher, as lenders seek to offset increased risk , further reducing access to credit for the small-scale farmer . Group-liability microfinance models, though popular in urban markets to reach low-income borrowers through social guarantees, may be ill-suited to serve smallholders in contexts where dominant risks driving default like weather and price shocks are common among members in the localized group. Group members will be unable to insure other members who cannot pay off a loan if, for example, everyone’s harvest is devastated by the same local flood or pest. On the demand side, demand from farmers for formal credit products is low. Even where formal financial products are available, farmers may opt to borrow money from within their social networks, or informal lenders. Preliminary findings from the rollout of Kshetriya Grameen Financial Services , a microfinance portfolio in rural Tamil Nadu,aeroponic tower garden system show that 72% of farmers’ loans at the beginning of the season are from formal sources, but only 35% are from formal sources by the end of the season.

Farmers seem to shift to informal borrowing given quick loan approvals and more flexible loan terms are available. This use of informal borrowing is particularly prevalent among marginalized farmers: 82% of the agricultural loans taken out by marginal farmers were from informal sources compared to 46% among medium-landholding farmers . Even where formal financial services are available, they are often highly disadvantageous to smallholder farmers. Farmers’ credit needs are different from urban micro-credit customers for which the common micro-credit products are designed, with weekly repayments and group liability. Most loan offers and repayment schedules are poorly timed to fit seasonal production cycles and price fluctuations. Uncertainty or risk aversion can also make farmers hesitant to take on debt. Profits in farming are uncertain, and are often low without complementary investments. Options for collateral to back a loan are limited in these environments, and assets like land may be too fundamental to basic livelihood to risk in order to access a line of credit , or unacceptable to back a loan in insecure contracting environments. Accessing and using financial products can also be even more difficult for farmers without high levels of financial literacy.technology adoption, pulling from a range of rigorous non-experimental work and some theoretical work to characterize the constraint facing farmers. We then summarize findings from recent randomized evaluations in an effort to distill policy-relevant insights.Agricultural income streams are characterized by large cash inflows once or twice a year that do not align well with specific times when farmers need access to capital to either make agricultural investments or, for example, pay school fees.

If there is limited access to credit in an area, farmers may not have cash on hand to make agricultural productivity investments unless they are able to save, or can afford the potentially high interest rates of informal lending. However, saving can be difficult for farmers given their limited resources, a variety of demands on their money, and the seasonal cycle of production and prices of their agricultural production. Credit and saving products could help farmers make investments in inputs and other technologies by making cash available when needed. Yet many developing countries, and particularly rural areas, have limited access to formal financial services that could provide this liquidity. Credit constraints have been reflected in farmers self-reports , and are associated with less use of productive inputs like high-yielding varieties . On the supply side, formal financial service providers are often unwilling or unable to serve smallholders. Few products suitable to agricultural livelihoods are available, and despite the wide proliferation of microfinance institutions , most are limited to non-agricultural activities given there are substantial challenges inherent in long-cycle agricultural lending . Lenders in these contexts charge high interest rates to help offset their assessment of the risk that loans will not be repaid. These higher interest rates can, perversely, have the effect of attracting only borrowers with no intention of repaying , thus driving interest rates even higher, as lenders seek to offset increased risk , further reducing access to credit for the small-scale farmer . Group-liability microfinance models, though popular in urban markets to reach low-income borrowers through social guarantees, may be ill-suited to serve smallholders in contexts where dominant risks driving default like weather and price shocks are common among members in the localized group. Group members will be unable to insure other members who cannot pay off a loan if, for example, everyone’s harvest is devastated by the same local flood or pest. On the demand side, demand from farmers for formal credit products is low.

Even where formal financial products are available, farmers may opt to borrow money from within their social networks, or informal lenders. Preliminary findings from the rollout of Kshetriya Grameen Financial Services , a microfinance portfolio in rural Tamil Nadu, show that 72% of farmers’ loans at the beginning of the season are from formal sources, but only 35% are from formal sources by the end of the season. Farmers seem to shift to informal borrowing given quick loan approvals and more flexible loan terms are available. This use of informal borrowing is particularly prevalent among marginalized farmers: 82% of the agricultural loans taken out by marginal farmers were from informal sources compared to 46% among medium-landholding farmers . Even where formal financial services are available, they are often highly disadvantageous to smallholder farmers. Farmers’ credit needs are different from urban micro-credit customers for which the common micro-credit products are designed, with weekly repayments and group liability. Most loan offers and repayment schedules are poorly timed to fit seasonal production cycles and price fluctuations. Uncertainty or risk aversion can also make farmers hesitant to take on debt. Profits in farming are uncertain, and are often low without complementary investments. Options for collateral to back a loan are limited in these environments,dutch bucket for sale and assets like land may be too fundamental to basic livelihood to risk in order to access a line of credit , or unacceptable to back a loan in insecure contracting environments. Accessing and using financial products can also be even more difficult for farmers without high levels of financial literacy.These credit market inefficiencies result in limited access to liquid capital from formal financial services. There is policy appetite to leverage new technologies and approaches to expand formal credit and savings mechanisms to rural households, particularly given the proliferation of micro-credit in urban markets. But even where micro-credit has expanded widely among low-income urban clientele, evidence from randomized impact evaluations show limited ability for micro-credit to transform the average entrepreneur’s business productivity and revenues, instead providing value through increased flexibility in how households “make money, consume, and invest” . In the smallholder context, we focus specifically on whether expanding access to formal credit on the margin of what is already available shows potential to unlock productive, profitable investments that improve rural livelihoods. Where expanding access to credit shows potential, studies investigate product designs aiming to increase credit access and their benefits specifically for smallholder farmers.Although the experimental evidence suggests that an injection of credit alone is unlikely sufficient to transform smallholders’ livelihoods, there is some encouraging evidence from approaches with careful product design. Financial service design innovation, particularly to encourage storage or savings, can generate more supportive services for farmers that can help them make investments or manage their volatile livelihoods. There is policy appetite to identify whether digital financial services will be able to connect rural borrowers to lending institutions and encourage financial behavior conducive to agricultural investment. More research is needed on these digital financial service channels and product designs, to understand their potential to support farmers’ financial portfolios in a manner that protects farmers while encouraging profitable investments. More research is needed to develop and test credit product designs and delivery channels that fit smallholders’ needs with respect to the timing of offers, repayment structures, and collateral agreements.Smallholder farmers have limited buffer stocks to cope with volatile food prices and climate uncertainty, and typically have few formal financial services to protect them from risk. The systemic risks of agricultural production jeopardize smallholder farmers’ ability to recoup their investments at harvest. Risk exposure therefore plays an important role in farmers’ agricultural investment decisions, including the use of productive inputs like fertilizer .

Rural communities have developed many informal mechanisms to cope with risk. For example, households may buy or sell assets in response to fluctuations in income , and communities may temporarily assist households experiencing a negative shock like an unexpected medical expense with the expectation that the household will do the same for others in the future . While these strategies are useful, in many cases they are insufficient. Farmers face many sources of uncertainty beyond weather and environmental factors including natural disasters, pests, and disease. Price risk and relationships with output markets can jeopardize farmers’ ability to recoup their investments at harvest, and such risks can depress productive input use. In addition to the risks inherent in the agricultural production status quo, new technologies often bear specific risks, such as uncertainty about how to use the technology correctly and how to market the output16. The classic economic view of poor farmers is that their lack of savings and other resources to fall back on causes them to prefer agricultural approaches with more reliable, but lower, average returns. Households often diversify their sources of income to spread around risk . Farmers may see the adoption of new technologies as risky, especially early in the adoption process when proper use and average yields are not well understood. Technologies that carry even a small risk of a loss may not be worth large expected gains if risks cannot be offset .So, while investments exist that could increase profitability, these may also increase the risks of farming. Behavioral biases also come into play around risky decisions . Risk averse farmers may prefer a more certain, but possibly lower, expected payoff over an uncertain payoff from unfamiliar technologies. Ambiguity aversion can lead farmers to stick to their status quo, preferring known risks with a more familiar probability of gains and losses, rather than unknown risks, even in cases where these choices may actually be less risky. Both risk and ambiguity aversion are important considerations when looking to encourage take-up of novel risk mitigating financial products or technologies . Evidence exists that rural households are able to mitigate idiosyncratic risk , but that rural residents are relatively unprotected against aggregate risks – weather and crop price shocks – common to smallholder rain-fed agriculture in poorly integrated markets . Given extreme weather events can destroy a large portion of harvest across a region, and that such weather events are only increasingly likely given global trends including climate change, there is a need for effective risk-mitigation strategies to protect farmers from these aggregate risks.“Linking credit with insurance has mixed results, suffering from the same demand problems that have beset standalone index insurance. The offering of indemnified loans that interlink an insurance product with credit appears promising, but demand for such loans has been shown to be surprisingly low in the few trials that have tested this mechanism ” . Linking credit with insurance has even been shown to drive down credit demand . Recent research has found that companies that engage in contract farming can be well-positioned to adjust the timing of insurance and payment arrangements to increase take-up. Casaburi and Willis find that when a large private company engaged in contract farming in Kenya offered to provide insurance to sugar cane producers by deducting premiums from farmer revenues at harvest time, take up rates at actuarially fair prices were 71.6%, 67 percentage points higher than the equivalent standardly timed contract.