Pronounced hygric seasonality determines the regional climate and, thus, the characteristics of rain-fed agriculture in the Peruvian Callejón de Huaylas (Cordillera Blanca). Peasants in the Cuenca Auqui on the eastern slopes above the city of Huaraz attribute recently experienced challenges in agricultural production mainly to perceived changes in precipitation patterns. Statistical analyses of daily precipitation records at nearby Recuay (1964 to 2013) and Huaraz (1996 to 2013) stations do not corroborate the perceived changes. Either insufficient temporal resolution of available precipitation records or other environmental and sociopolitical factors impacting traditional farming methods may be the reason for the lack of concordance between the two information sources investigated in this study.
Scientific evidence of climate warming and of projected resulting impacts can provide the basis for a responsible and efficient adaptation strategy if implemented in a timely and careful fashion, but can also be misused to legitimize particular interests (Arnall et al., 2014; Dietz, 2011; Neuburger, 2008). While the physical aspects of climate change are, though complex, of relatively straightforward nature, societal processes in reaction to them are contingent upon and characterized by the different interests, positions and vulnerabilities of affected groups (Postigo et al., 2008; Sietz, 2014; Zimmerer, 1993).
A region of specific interest is the Callejón de Huaylas (the valley drained by the Río Santa) in Peru, where water availability is determined by particular climate and topographical settings (e.g., Kaser et al., 2003). While the tropical atmosphere is thermally homogeneous, the region is characterized by single-peaked hygric seasonality. Precipitation increases from August towards the October to April core wet season and is close to nil during June and July (e.g., Bury et al., 2010; Kaser and Osmaston, 2002; Mark et al., 2010; Schauwecker et al., 2014). Dry season runoff, and thus water supply, is comprised of up to two thirds glacial melt water from the Cordillera Blanca (e.g., Baraer et al., 2012; Kaser et al., 2010; Mark and Seltzer, 2003, 2005). They smooth the seasonal runoff to a degree that varies with the proportion of sub-catchments that are covered by glaciers (e.g., Kaser et al., 2003; Mark and Seltzer, 2003). While the highest glacier cover of up to 41 % is found in the northern Cordillera Blanca valleys, rivers draining the western Cordillera Negra are lacking in glacier contribution (e.g., Kaser et al., 2003).
Glacier contribution definitely has a considerable effect on the runoff of the Río Santa during the dry season (Bury et al., 2013; Carey et al., 2014) and even more so on the tributaries draining the Cordillera Blanca. Both ancient and modern channel systems have witnessed the sophisticated use of river water for agriculture and other needs (Bury et al., 2013; Gelles, 2001). Many studies were dedicated to the impact of glaciers on runoff and water availability in the region (Baraer et al., 2012; Carey et al., 2014; Mark et al., 2010). The increasing knowledge of human-caused climate warming and resulting impacts has attracted much attention in the region up to now (Baraer et al., 2012; Bury et al., 2013; Carey, 2010; Carey et al., 2014; Chevallier et al., 2011; Juen et al., 2007; Mark et al., 2010; Vuille et al., 2008) and, among other interests, our interdisciplinary research team also focuses on this issue.
Yet, by gradually deciphering natural components of water availability, societal practices of water use, and emerging trends of water conflicts, we have identified important non-glacial aspects of the question of water supply in our study region. Even on the slopes of the heavily glaciated Cordillera Blanca many small-scale farmers have no access to nearby glacier-fed river runoff due to unequal land and water distribution systems.
Accounts from local peasants suggest that changes in precipitation patterns, e.g., during the onset of the wetter season (August–September), the traditional period for ground preparation and first seeding, have caused detrimental effects on the crops' growth and, thus, on overall agricultural production. However, human perception can often fail to accurately determine the drivers of concern (e.g., precipitation, solar radiation, temperature, deforestation, changes in seed types etc.) for the experienced impacts (changing soil moisture, problems with seedlings and harvests). A full range analysis of crop yield and precipitation data, of forest degradation, soil erosion, changing seeds, cultivating methods, development and/or liberalization of agricultural markets, political programs and dominant discourses etc. would lead to a most comprehensive answer to whether and why crop yields may have changed and how counter measures could be applied. Yet sustainability farming is rarely accompanied by systematic data collection and change monitoring, hindering a comprehensive analysis of drivers of alterations in crop growing. General statistics on agricultural production changes for the entire Department of Ancash do not allow for the derivation of local information and understanding the complexities of changes (Bury et al., 2013).
Also, precipitation data in the Callejón de Huaylas have been recorded more from the perspective of hydropower use than of agriculture, and thus long-term measurements with high temporal resolution (at least daily values) as required for analyzing potential impacts on crop yields are rare. Nevertheless, for this study we were able to assemble time series of daily precipitation totals for two sites in the Southern Cordillera Blanca, for the periods 1964 to 2013 and 1996 to 2013 respectively. These data allow us to examine one potentially powerful (and the most blamed) driver of the experienced changes in rain-fed agriculture.
In this study we examine the issue by (i) characterizing agricultural practices of Andean peasant families along the Río Auqui (crop types used, seasonal cycle of sowing, growing and harvesting) and presenting and evaluating the peasants accounts of changes, (ii) analyzing available information on local precipitation, and (iii) touching on potential effects on small-scale farming as far as possible from available data. Aspects of climate change impacts not investigated in this study and potential other disturbances of agricultural performance will be briefly referred to in the discussion section.
The outcome of this study may shed light on other peasant communities in the region whose economy is based on rain-fed agriculture and, more generally, on mountain regions with similarly vulnerable communities and with similarly poor availability of information. It also takes account of potential complications caused by the different approaches of scientific groups in an interdisciplinary setting and by bringing together epistemologies represented by Western scientific knowledge with peasants' local knowledge (e.g., Boelens, 2014; Escobar, 2008; Klein et al., 2014; Lennox and Gowdy, 2014). We emphasize that this study only concentrates on one variable (precipitation) out of a series of potentially interrelated variables explaining perceived changes. It will, therefore, rather point out open questions then provide conclusive answers to the complex problem. In fact we consider the outcome of our methodical experiments at the interface of the, by their nature, explicitly different knowledge systems as a major result. We therefore provide the details of the approaches taken from both sides for providing insight to readers from different scientific disciplines.
Our study site (here called Cuenca Auqui) stretches from the city of Huaraz along the slopes south of the Río Auqui up to the highest settlements close to Río Shallap and includes five main villages with about 1500 inhabitants in total: Los Pinos, Ichoca, Collyur, Paquishka and Jancu (Fig. 1). The narrow bottom of the valley is well-defined by steep slopes reaching altitudes up to 4500 m a.s.l. that become gentler towards the crests. Aside from some houses at the valley flanks, all settlements are located close to the road at an altitude of 3200 m a.s.l. (Los Pinos) to 3800 m a.s.l. (Jancu). The cultivation area in the Cuenca Auqui is naturally concentrated towards the valley bottom but also extends to the adjacent slopes. Irrigation is currently only available for relatively small areas close to the river (Fig. 1). The irrigation channel along the upper slopes has never been in operation yet. Since 2014 it is under reconstruction to improve the urban water supply. From a hydrological viewpoint the Cuenca Auqui stretches from the Río Santa into the heavily glaciated Cordillera Blanca, which together with the ice-free Cordillera Negra in the West, defines the Callejón de Huaylas.
The study site (Cuenca Auqui) within the Rio Santa valley in Northwestern Peru.
With the implementation of the agrarian reform in 1969, four former
haciendas (colonial large-scale farms) at the southern side of the Río
Auqui were divided into small plots and distributed among the local farmers
of the Cuenca Auqui. Since then, agricultural activities are characterized
by subsistence production of potatoes, grain and corn, with surpluses being
sold at the markets in Huaraz. Fast-growing eucalyptus has been planted for
construction, heating and cooking purposes. An irrigation channel fed by the
Río Auqui could supply most farmland in the watershed, but it is out
of service and the water is contaminated by heavy metals Personal
communication with Instituto de Montaña, Huaraz, July 2015.
Based on the idea that local people are closely linked to their environment through continued practice of resource-based livelihoods, we collected information on the local ecological knowledge of peasants in the Cuenca Auqui as a first step. Emphasis was on local climate and environmental changes, with particular interest in agriculture and related community activities (e.g., Agrawal, 1995; Alexander et al., 2011; Klein et al., 2014). Since the memories of individuals are limited in space and time they are not able to mirror the complete natural and societal processes. We understand this type of human knowledge as a subject of continuous iteration between individual and collective perceptions, practices and beliefs, modified by specific socio-political and discursive dynamics (Boillat and Berkes, 2013; Orlove and Caton, 2010; Orlove et al., 2008; Zimmerer, 2010, 2011). Accordingly, the derived information represents a snapshot of the broad and highly complex local knowledge about environment, society and history.
List of interviews of farmers in the communities of Cuenca Auqui.
We conducted semi-structured and narrative interviews in all five
communities of the Cuenca with peasant individuals and families of different
gender and in different stages of life (Table 1). Because of traditional
gender roles which define public discussions about water management and crop
production as a masculine-dominated sphere a greater proportion of
interviews was made with male community members. As most women are
exclusively involved in reproductive work in the household and in animal
husbandry, they referred to their husbands as “experts” in this field when
being asked about agricultural production. The interviewees were selected by
“snowball sampling” (Goodman, 2010; Heckathorn, 2011)
starting with the community authority, who then indicated other families in
their community. A sequence of interviews was conducted in each community
until reaching saturation to ensure that no new themes emerged. We conducted
the interviews in Spanish and were supported by a local translator when
interviewees only spoke Quechua. All interviews included questions about
household, agricultural practices (products, technology and intermediate
goods, man power, agricultural calendar), community life, and questions of
integrating environmental character. Main focus was on their experienced
changes for each issue as well as the mutual dynamics over the last decades.
Despite the frequently used vague time references like “in former times”
or “before the earthquake” In 1970 an earthquake caused huge
damages in the region (Lipton, 2014).
In a second step we applied additional methods of qualitative analysis. We conducted expert interviews with the community authorities, the elected political representatives of each community. They represent the community externally and internally, coordinate community activities such as maintenance of communitarian infrastructure like roads, irrigation channels, water reservoirs, and community centers, and settle disputes within the community. Additionally, we questioned officials of the Juntas Administradoras de Servicios de Saneamiento (JASS, local administrative boards of sanitation) and Juntas de Riego (committees of irrigation) that are responsible for water supply and irrigation in the communities. To capture most recent discussions on potential climate related changes we organized a participative mapping meeting in December 2013 with 16 representatives of all communities in the Cuenca. The representatives designed maps of their communities showing relevant issues and changes related to climate, agriculture, water resources and community life. A participants' comparative discussion revealed similarities as well as differences between communities in the Cuenca Auqui.
In a final step we extended our interviews to individuals and institutional experts outside the Cuenca Auqui in order to relate our knowledge to the wider upper Callejón de Huaylas. We therefore conducted “go-along interviews” (e.g., Anderson, 2004; Bergeron et al., 2014; Evans and Jones, 2011) with two informants from neighboring communities. One of them is a local guide from Llupa, who regularly accompanies international scientific expeditions in the Cordillera Blanca, and the other is a local historian of the community Chontayoc, located in the Cordillera Negra. These interviews yielded details about ecological conditions and agricultural practices in the Cuenca Auqui and the nearby Río Santa valley. Furthermore, we interviewed 26 representatives of public institutions and NGOs (non-governmental organizations) in Huaraz which deal with agricultural and environmental issues. From these “expert interviews” we gathered technical, agronomic and political information about structures and dynamics in agriculture, water policies, population, and migration at regional level. With these interviews we cross-checked the Cuenca Auqui peasants' reports and added details.
All interviews and meetings were recorded, transcribed and analyzed with the software MaxQDA. In the digital documents we marked all comments on agricultural practices (including experienced changes), on environmental and climate issues linked with agriculture, and on all connections established between changes in climatologic phenomenon and agriculture.
To enhance the reliability of our results, we only include individual statements which were confirmed in focus group discussions or by institutional or NGO representatives in our analysis. Since individual perceptions and collective memory in the Cuenca Auqui and beyond are mutually linked, only very few statements differ from the general view.
Agricultural calendar of the main crops used in the Cuenca Auqui.
The peasant families of the Río Auqui watershed cultivate an average area of around 3 hectares per family which are distributed in small plots over different altitudes of the valley (Fig. 1) in order to guarantee diversified production for each family (Sietz et al., 2012; Vos, 2010; Zimmerer, 2011). If possible, families combine irrigation and rain-fed agriculture, but overall only few are privileged in having access to irrigation for year-round cultivation. The large majority of the families depend entirely on rain-fed agriculture and, consequently, on precipitation. The cultivation calendar in Fig. 2 results from our interviews and fieldwork.
Rain-fed agriculture is strongly dominated by the pronounced seasonal cycle in precipitation and crops are vulnerable to changes during different phases of the cultivating cycle. Different crops and the different altitudes and climates in which they are cultivated increase the resilience of a community, yet irregularities or extremes during specific times of the agricultural year are still viewed with concern.
The first rain events after the core dry season in August and September are
of particular importance as they mark the start of the rainy season.
According to the reports, these first rainfall events are of gentle
character, providing favorable conditions for preparing the fields. They are
of great importance to agricultural life and were celebrated with festivals
following ancient traditions. The enduring importance of these gentle rains
is evident by the persistent use of the ancient
The main crops for subsistence in the Cuenca Auqui are potato, wheat, corn,
and the traditional
Besides some high altitude adapted species of wheat, corn and wheat are mainly cultivated in lower areas (below 3500 m a.s.l.) because they are vulnerable to frost. While corn is sowed in August or September and has a relatively long vegetation period of approximately 7 months, wheat is sowed in December and harvested between June and July. Both crops consume a lot of water and are vulnerable to dry spells in their growing period, as well as to frost and hail toward the later stages of growth. Furthermore, they are sensitive to wet conditions and heavy rain events in the ripening period.
For
Over a period referred to as the “last decade(s)”, as compared to undated
“past times”, farmers report the following changes:
The The beginning, duration, and end of the wet and dry seasons have become more
variable and, in general, rainfall has become more irregular, which
complicates successful farming overall. The occurrences of hail and heavy rain events have become more frequent
during September and October, when corn and potato are in their sensitive
phase of germination and initial growth, but also throughout the entire wet
season, causing high surface runoff and increased soil erosion. Damages to
crops during both flowering and the harvest season are more frequent. Ground frost has become more frequent during September and October, damaging
the crops in the early vegetation period.
The applied methods (narrative interviews and reports from group meetings)
do not allow for strictly categorizing the obtained information. Thus,
quantified analyses are not possible but the statements clearly converge
among the communities. Our findings also mirror widely those found in an
earlier study in the region (e.g., Mark et al.,
2010) where the farmers perceived similar changes in the precipitation and
weather patterns.
On a group to group level differences in the perceptions depend on both the location and altitude of the communities' plots as well as on the water demand characteristics of the planted crop types. Families in the higher altitude communities of Jancu and Paquishka mainly cultivate traditional crops which are relatively resistant to heavy rains and dry spells. They identify ground frosts as the biggest challenge. In turn, communities at lower elevations such as Los Pinos, Ichoca and Collyur plant mainly modern crop types on relatively steep slopes. They feel most challenged by changing precipitation variability as well as increased heavy rainfall frequency that leads to soil erosion.
Climate change is mainly seen in view of environmental justice (Schlosberg, 2007) with causes in both the industrialization in “the First World” on a global and air pollution from mining as well as air and car traffic on the regional scale. Climate change consequences are sensed as a burden without having benefits of modernization and wealth.
Most questions related to changes in rain-fed agriculture require at least daily temporal resolution. Only two stations in the surroundings of our study area (Fig. 1) provide daily precipitation values (07:00 to 07:00 LT) over time periods of an appropriate length. Huaraz at the bottom of our study area (3052 m a.s.l.) has a record of daily precipitation from 1996 to 2013 (with several gaps) when merging the station records of “Huaraz” and “Santiago Antunez de Mayolo”. Recuay at 3445 m a.s.l. is about 25 km up-valley along the Rio Santa from Huaraz and covers a much longer period from 1964 to 2013, with gaps of 156 days in total that could be closed with data from Recuay Sut and Laguna Ututo, 1 and 10 km from the Recuay station respectively. The data were made available by SENAMHI, National Meteorological and Hydrological Service of Peru.
In order to test the representativeness of these two stations for the study
area we were able to use unpublished time series of weekly precipitation
sums measured at Llupa (3435 m a.s.l.; from 2003–2013) relating to
research projects conducted by our group in the Cordillera Blanca. Over
approximately 10 years of overlapping time series, mean weekly precipitation
deviated by only 2 mm week
The farmers' reports and concerns reflect the strong influence of several
features in the annual precipitation cycle on farmers' lives and the
agricultural year in the Cuenca Auqui. The steadiness of these
characteristics determines the success or failure of sowing, growing and
harvesting (Ambrosino et al., 2014; Kniveton
et al., 2009; Raes et al., 2004). To extract the agriculturally relevant
information from the seasonal cycles of daily precipitation to be compared
with the farmers' experiences, we defined eight criteria, mainly empirically and
inspired by methods presented, for example, by Laux et al. (2008).
In the following,
Precipitation features derived from daily precipitation sums based
on measurements in Recuay (1964 to 2013 with gaps) and Huaraz (1996 to
2013), with the criteria described in Sect. 3.3. The
Time series of
Onset day wet season: First sowing conditions after 1 August:
sum( sum( Dry spell during wet season: sum( Heavy precipitation day: Onset day of the dry season: Wet spell during dry season: sum(
For a small number of years with unusual precipitation patterns, e.g., our
criteria 2 and 6 did not yield reasonable results (dry season onset Recuay
in 1985, wet season onset Recuay 1972) or even failed (wet season onset
Recuay in 1992). We accept these minor problems as we tried to keep all
eight criteria as simple as possible to facilitate comprehensibility. More
sophisticated criteria could tend to over-interpret our limited information.
Some other missing values in individual years (e.g., 1974–1978 in Recuay;
2012 in Huaraz) are the result of data gaps in the precipitation records.
Different to the other criteria, criteria 3a and 3b, yielding start dates
for the sowing season, are based on information from literature as these
criteria are more objectively assessable than, for example, the
human-perceived onset of the wet season. Criterion 3a follows data presented
in Table 1 in Sanabria et al. (2014) which is the only study
we know that presents typical precipitation values required for planting of
different crop types in the region. Three days of consecutive precipitation
with total precipitation
For analyzing potential trends in the calculated features of the
precipitation time series we applied the Mann–Kendall trend test
(significance threshold set to the 90 % confidence level) to all features
presented in Figs. 3 and 4 for the 1981 to 2010 time span As for
the year 1991 no onset date for the wet season could be calculated
for Recuay, we had to ignore this year for the trend analysis of
the onset dates of the wet season in the period 1981–2010.
In order to facilitate a comparison between human perceptions and memories and measured records, we classified the precipitation data along the eight criteria presented in Sect. 4.2. The results are presented in Fig. 3 by starting with the agricultural year on 1 August. In the following sections a value given for a certain year (e.g., 2003) refers to the agricultural year (1 August 2003 to 31 July 2004). Data statistics are discussed along the issues raised by the farmers and listed in Sect. 3.3.
We first comment on the
As a first objective indicator of the start of agricultural year we present
the onset of the wet season which should, by definition (Sect. 4.2),
approximate the time when the weather conditions change from continuously
dry to frequently humid. For Recuay (1965–2012), the date is typically in
September or early October with the earliest onset calculated for 13 August (1988)
and the latest for 20 November (1972), the
arithmetic mean value being 23 September. For 1991, no onset of the
wet season could be calculated as there was an unusual dry period from
autumn 1991 to spring 1992 (also described by Schauwecker et al., 2014). For Huaraz
(1996–2012) the onset day occurred typically 5 days later than in Recuay
with a mean value centered on 28 September, the earliest day being
4 September (2001), the latest 25 October (2011). The mean
variability of the onset was
Analyzing the precipitation records in view of favorable sowing conditions
shows that criteria 3a and 3b (see Sect. 4.2) are typically met between
mid-September and mid-October at both measuring sites, with individual dates
ranging from 30 September
However, there are also cases where both criteria generated the same dates for good sowing conditions (e.g., 1999 and 2003) which were then followed by pronounced dry spells. Such patterns indicate particularly challenging conditions for farmers if, motivated by the first rainfalls, they sowed before the likely harmful dry spells. A rough estimate suggests that such potentially problematic conditions occurred in 7 out of 17 years between 1996 and 2012, both in Huaraz and Recuay. As a side note it is worth mentioning that on two occasions (1971 and 1989) first sowing conditions (criterion 3a) occurred during a (rare) wet spell in August (earliest date in record: 14 August 1989) which could also have provoked farmers to sow, only to face a pronounced dry period soon after.
Dry spells also occur from December to April, reflecting pronounced variability of precipitation even during the core wet season. However, dry spells during the middle of the wet season were less frequent than during the transition periods (expect for few years like 2000 at both sites and the exceptional year 1991 in Recuay). They also have lower potential to harm the plants, being in advanced stages of growth by then. Overall, as visible in Fig. 4c, we have no evidence for increased frequency of dry spells in each agricultural year (no significant trend in Recuay data). Also the mean and maximum length of the dry periods (consecutive dry spells) lack significant trends (Fig. A1).
Heavy precipitation days (here defined as
The transition from wet to dry conditions marked by the onset day of the dry
season (criterion 2) towards the end of the agricultural year is centered
around 17 May (earliest on 23 April in 1981; latest on 14 July,
in 1984, possibly an outlier caused by our algorithm as a
consequence of unusually high precipitation values of around 10 mm within
3 days at the beginning of July). For Huaraz the dry season onset is found to
be around 16 May, earliest on 23 April in 2005 and latest on
26 June in 1999, the latter again as a consequence of unusually high
precipitation during the preceding 3 days (
As with the onset of the wet season, there is quite a high year-to-year
variability in the calculated onset dates of the dry season at both sites
(
The overall length of the wet season is plotted in Fig. 4a and shows a
year-to-year variability between
Mean total precipitation during the wet season was 810 mm for both sites with
a pronounced variability between
Finally, we also tested the time series of monthly precipitation (Fig. A3)
with the Mann–Kendall analysis at 90 % confidence level for possible
trends in Recuay between 1981 and 2010. As shown in Fig. A3 (dashed red
line) March precipitation increased significantly by approximately
Multiple environmental changes are perceived by peasants living on the eastern slopes above the city of Huaraz in the upper Callejón de Huyalas. The most prominent changes – as expressed in interviews collected for this and for a former study (Mark et al., 2010) – were felt in the context of climate, such as the shrinkage of glaciers, decreasing dry season river discharge, or changes in weather patterns. These reports stimulate hypotheses to be tested against measured records. Whereas Mark et al. (2010) intensively investigated changes in (glacier-fed) river runoff, we here focused on temporal precipitation patterns in view of their impact on rain-fed agriculture.
Daily time series of precipitation yield interesting insights in rainfall
characteristics of the last decades and allow the comparison with peasants'
statements: starting in search of the light
Nevertheless, data presented in Fig. 2 show that the agriculturally relevant sowing period for potatoes typically starts in September and continues until mid of October. For most years the calculated onset date of the wet season and the dates for the first sowing conditions after the dry season fell into that period (Fig. 3). Yet, the pronounced year-to-year variability of these dates challenges agricultural success, especially in the absence of reliable precipitation forecasts. Furthermore, it is hardly possible to predict devastating dry spells following a couple of days or weeks with good sowing conditions. Our “hind-cast” detected several such potentially harming sequences (Fig. 3) even though we could not find long term changes in the dry spell frequency (Fig. 4c).
Overall, and despite no detectable trends in the total amount of precipitation during the wet seasons (Fig. 4b) nor any other trend, the high inter-annual variability of (1) the timing of the onset of the agricultural year (as determined by the first pronounced precipitation event) and (2) dry spells during the wet season, especially during the very sensible early phase of plant growing, kept rain-fed farming constantly challenging and likely favored perceptions of water scarcity (Murtinho et al., 2013).
In the absence of adequate temporal data resolution, this study cannot give conclusive answers on the potential impacts of intense precipitation events, possibly accompanied by destructive hail and flooding. For daily rainfall sums we found neither a trend in the frequency of heavy precipitation days during the wet season, nor during September and October, the period recognized as most sensitive by the peasants.
We have not investigated thermal conditions but the perceived increase in the frequency of ground frosts in the early growing season (as stated by some farmers) contradicts increasing (minimum) temperatures as reported by Schauwecker et al. (2014) or in Vuille et al. (2015). The increases in minimum temperatures are reported to be most pronounced in the dry and early wet season. Also freezing level altitudes were rising during the last decades according to studies of Bradley et al. (2009) and Rabatel et al. (2013).
To extend information about climate impacts on rain-fed agriculture beyond the results presented in this study, it would be desirable to analyze local extremes in temperature and also precipitation intensities based on data collected by onsite automatic weather stations with high temporal resolution in upcoming studies.
Potential impacts of future climate conditions are currently highly uncertain due to missing or unreliable data of high spatial and temporal resolution as required for investigating impacts on rain-fed agriculture in the complex Andean terrain (Sanabria et al., 2014). Further uncertainty is due to the questionable future evolution of the El Niño Southern Oscillation (e.g., Vecchi and Wittenberg, 2010) which affects the year-to-year climate variability in the Cordillera Blanca (e.g., Garreaud et al., 2009; Vuille et al., 2008).
Beyond climate, there are several other factors not investigated within this study that could have impacts on the small-scale rain-fed farming in the study region with the potential to explain the perceived water scarcity. As peasants' reports indicate, neoliberal agrarian policies since the 1990s and the loss of manpower due to emigration, particularly of young community members (Crabtree, 2002; Lynch, 2012; Trivelli et al., 2009) are among the potential causes for the decrease in the traditional adaptive capacity of peasants in the Cuenca Auqui. Deforestation, the cultivation of water-demanding eucalyptus trees, land use change followed by soil erosion, and the change from traditional to industrial seed types are some of the manifestations of the manifold changes.
Independent from climate change, socioeconomic and ecological changes have presumably challenged rain-fed agro-production considerably. Partially because being out of the scope of our research project but also because of the lack of comprehensive information about, e.g., land use, agricultural methods, differences in income opportunities between rural and urban areas etc., we are not able to give conclusive answers in the light of the full complexity of the issue. In fact, we speculate that, because of the missing assessments on impacts of ecological and societal evolutions, climate change is currently seen as a “clear” reason for the increasing difficulties for cultivation. The perception is a blending of both individual and community sensed experiences mirroring the global discourses on climate change as regionally reproduced by NGOs, governmental institutions, and international development agencies.
The reason for the converging responses within each group (Sect. 3.3) is most likely the result of collective knowledge production on changes in weather and climate patterns, intertwined with dominant global discourses on climate change. First explorative analyses of interviews with all relevant stakeholders in the region indicate a strong linkage between local and global discourses. The positioning of small holder families within the climate change discourse may represent a strategy to be heard within the society and to benefit from climate-related political measures. Similar findings of strong interrelations between peasant perceptions, collective memory, dominant discourses and specific agricultural practices were found in other rural environments in the Andes with similar socio-ecological connections (Postigo et al., 2008; Sietz et al., 2012; Zimmerer, 1993, 2011). In order to refine and specify the multiple interactions, detailed analyses of the dynamics of climate and environmental change discourses would be needed.
We investigated agricultural practices and peasants' perceptions about
climate impacts on rain-fed farming in small settlements of the Cuenca Auqui
in the Cordillera Blanca, Peru, and compared these with agro-relevant
precipitation features derived from daily data recorded at neighboring
stations. The most important agricultural crops are potatoes (native and
industrial varieties),
Farmers and local experts concur in their statements that changes in the climatic conditions have detrimental effects on agriculture. This also corresponds generally to findings made by Mark et al. (2010). Overall they view rain-fed agriculture as having become more challenging in recent years/decades and believe the reasons are changed precipitation patterns with less rain in August and the early wet season, more variable onset dates and durations of wet and dry seasons, and more intense rainfall events. Increased frequency of temperature-related ground frost was also reported.
Our precipitation analysis cannot confirm any precipitation changes but show high year-to-year variability in the onset dates of the wet season, the dates for the first sowing conditions after the dry season, and the number of heavy precipitation events per agricultural year. We also found that in several years pronounced dry spells occurred shortly after several wet days in the early cultivation season, encouraging the farmers to sow too early. Generally, high variability in rainfall has been shown to provoke perceptions of water scarcity in other Andean regions (Murtinho et al., 2013).
In conclusion, the year-to-year variability in seasonal and total precipitation during the agricultural year generally poses challenges for successful rain-fed farming in the region but no trends at all can be seen in the available precipitation data. Potential effects of heavy precipitation events and trends in their frequency could only partially be addressed in this study due to the lack of adequate data.
The study has also shown both the challenges of interdisciplinary research on complex climate change impact issues and the strong need for further developing scientific approaches that analyze all factors of concern: environmental as well as socio-political factors and their interconnectedness. The present study can only exclude precipitation changes as a likely reason for a perceived water scarcity. Precipitation information at higher spatio-temporal resolution and a series of other potential factors including factor-combinations need to be looked at for a holistic analysis of pressures on the small-scale rain-fed farming in the study region.
Annual precipitation sums in
Monthly precipitation totals from the Recuay and Huaraz time
series. The number in the
W. Gurgiser performed the precipitation analysis, and coordinated and merged the contributions. I. Juen prepared the precipitation data and assisted in preparing the manuscript. K. Singer carried out the interviews and the related analysis. S. Schauwecker assisted in preparing the precipitation data. M. Hofer advised precipitation downscaling attempts. W. Gurgiser, M. Neuburger and G. Kaser wrote the paper. All authors continuously discussed the methods and results, and developed the study further.
This study was funded by an interdisciplinary DACH project of the Austrian Science Fund (FWF), project number I900-N21, and the Deutsche Forschungsgemeinschaft (DFG), project number NE 903/4-1. We thank three anonymous reviewers for their comments on an earlier version of the manuscript that led to substantial improvements of this paper. We thank Leona Faulstich, Jana Lüdemann, Nina Scheer and Alexander Döpke for their support during field work and for figure preparation. Thanks are due to SENAMHI, National Meteorological and Hydrological Service of Peru for sharing information and measurement data. We are grateful to David Parkes who checked and improved the language. Edited by: S. B. Roy