5.1. Los Mochicas del Norte Valleys
This area includes the valleys between Rio La Leche and Rio Jequetepeque and has an ethno-historical significance [
131]. The earliest work identified by citation database queries is the one of Craig and Shimada [
16] on the Batan Grande archaeological complex (see
Figure 1 for location). Recent Quaternary stratigraphy, analysed along a modern hydraulic excavation, suggests to the authors that few deposits survive from the 1925/1926 El Niño event, except for slack-water beds roughly dated between 650 and 1000 CE by associated funerary pottery. Such deposits accumulate in areas of reduced velocity during flood flows and may be related to episodes of fluvial morphological adjustment and reshaping channel morphology [
66]. Unluckily, the contained archaeological finds do not allow more precise dating, and thus the use of the above deposit as a horizon marker is prevented. Such a question may be addressed by absolute dating of several samples.
South of Batan Grande, along the Jequetepeque Valley (
Figure 1), Dillehay et al. [
94] documented several sediment release signatures of slack-water deposits containing Late Moche and Chimu ceramics and
14C dated between 415 and 1420 CE. These authors also documented different debris flow deposits as well as erosional truncation of floors at several archaeological sites, all likely associated with El Niño events. The multidisciplinary character of this study and the numerous radiocarbon dating of organic material extracted from alluvial deposits make the findings reliable. A long-term landscape shaping due to severe ENSO-related floods characterized the northern coastal desert before the arrival of the Spanish conquistadors. The human responses to the destructive effects of El Niño events are evaluated as “
highly sophisticated” by the authors. Large rebuilding activities on damaged hydraulic structures, inferred by stratigraphic analysis at different archaeological sites, seem in fact to have allowed communities to avoid socio-economic repercussions of the hydrogeomorphic calamities. Similar results are obtained by Huckleberry et al. [
113] for the inter-valley canal system named Racamuri (upper La Leche Valley,
Figure 1), a millennial construction that reached its maximum extension between 900 and 1470 CE. Despite ENSO-driven floods and droughts, such a hydraulic structure would continue to work for centuries, likely even after the beginning of the colonial occupation. As a whole, for the northernmost valleys of coastal Peru, the selected literature data do not suggest drastic landscape changes or dramatic human responses to strong hydrogeomorphic episodes. Only the ca. 775 CE episode of abandonment of the Pampa Grande site (45 km southeast of Batan Grande) may be associated with a strong El Niño with good confidence [
129]. It must be highlighted that this date is located within the time interval of the event postulated in Ref. [
16]. Thus, a detailed geoarchaeological study on this case is suggested.
5.2. Huacas de Moche (Lower Moche Valley)
The settlement of Huacas de Moche is an early capital of the Moche state. It is located in the lower Moche Valley and includes the well-known buildings named Huaca del Sol and Huaca de la Luna (
Figure 2). In their pioneering works, Nials et al. (a) [
14] report “
the discovery of an El Niño catastrophe of a magnitude far greater and more devastating than all other such natural disasters striking the coast since the conquistadors first arrived in 1532. Transpiring about 1100 A.D., this prehistoric Niño was of unprecedented magnitude, and the devastation it wrought taxes the imagination of geologists and archaeologists alike”. Since about 1000 CE, the Chimu culture had completely transformed the desert landscape, building a complicated system of reclamation channels over an area of tens of square kilometers. According to the authors (who have surveyed hundreds of kilometers of ancient waterways), all the channels experienced massive erosion and dropped from use before the beginning of the successive sub-cultural phase (a few tens of years later). This allowed an approximate dating of the hydrogeomorphic disaster, which was consequently called the Chimu Flood. Later, much of the reclaimed area would revert to desert in short time. However, the lack of absolute dating makes this reconstruction questionable.
The water level of the Chimu Flood would be still recognizable on the western side of the adobe brick pyramid
Huaca del Sol (see Figure 7 on page 12 of Nials et al. (a) [
14]), marked by a notch about 10–15 m above the present Rio Moche level. However, according to the geomorphological reconstruction of the authors, the major landscape change happened at the middle Moche Valley, 5–10 km upstream Huaca del Sol, with the lowering of the alluvial plain due to erosion estimated in ca. 5 m [
15] (see
Section 5.3). Moreover, the authors stated that “
a very conservative estimate would be flood waters at least 2 to 4 times the size of the 1925[/1926] floods, the worst in the last 400 years” [
15].
Several works selected from Scopus and WoS databases by the method described in
Section 3 cite the works of the research group of Nials. Craig and Shimada [
16] hint at a possible regional correlation with the slack-water deposit they found along the Rio La Leche Valley (see
Section 5.1), thus laying the groundwork for the conceptualization of an 11th century El Niño event. Wells [
17] suggests the possible coincidence of the Chimu Flood with the so-called Naymlap Flood, an ethno-historically recorded hydrological disaster associated with the name of a cultural hero [
7,
132]. Again, Wells and Noller [
88] use the Chimu Flood to explicate the recurrence interval of the “Mega” El Niños in northern coastal Peru. Finally, Van Buren [
91], Nesbitt [
120], Huckleberry et al. [
113], and Caramanica et al. [
126] cite Nials et al. [
14,
15], simply to describe the destructive effects of the major El Niño events on the ancient irrigation systems and the consequences on the human communities. None of these studies deals with the reliability of the reconstruction provided by Nials et al. [
14,
15].
By archaeological excavations on
Huaca de la Luna, Uceda and Canziani Amico [
86] argued for the occurrence of moderate El Niño events before the Chimu Flood. They interpreted three layers of sediment with remnants of painted washed-off walls, found on successive floors of the temple, as evidence of intense precipitation and runoff processes, the most recent dated around 600 CE. This latter should have produced significant flow within the braided streams of the Moche Valley without, however, causing either changes in the landscape or the abandonment of the settlement as asserted by Moseley and Richardson [
83]. Instead, according to these authors, between 500 and 600 CE, “
flood water brought by El Niño struck the Moche capital, they leveled much of the city, stripping as much as 15 feet” (about 4.6 m) “
off some areas. It is unclear if the magnitude of destruction reflects more than one El Niño events, perhaps exacerbated by an earthquake. The survivors rebuilt their city only to see it gradually inundated by sand dunes that swept inland after forming on the beach at the mouth of the Moche River”. It is apparent how the interpretation of geoarchaeological data coming from the same site can even lead to conflicting hypotheses [
128]. Thus, the impact of the El Niño event around 600 CE on settlement and landscape at Huacas de Moche must be further investigated, especially with the aim of filling possible gaps in knowledge and addressing the above question.
The study by Prieto et al. [
133] on a mass-sacrifice event discovered at a site of the lower Moche Valley, in addition to reporting a possible landscape process due to an El Niño event dated between 1400 and 1450 CE, is a remarkable example of the indissoluble connection between geological and archaeological data in the late pre-Hispanic coastal Peru. According to these authors, “
stratigraphic evidences suggests that the sacrifice was made following a heavy rain/flood that deposited a layer of mud on top of the clean sand in which the children and camelids were buried”. The conspicuous number of
14C dated samples ensures a high reliability of the chronological attribution.
5.3. Caballo Muerto and Quebrada de los Chinos (Middle Moche Valley)
Caballo Muerto Archaeological Complex and Quebrada de los Chinos are located in the middle valley of Rio Moche (
Figure 1). The first includes the building named Huaca Cortada, from which data on the effects of El Niño events on settlement and landscape have recently been collected by Nesbitt [
120]. According to this author, the whole archaeological complex “
is situated within an environment susceptible to El Niño flooding”. Due to its geomorphological setting, especially the alluvial fan located 10.5 km north-east of Huacas de Moche (
Figure 3) may preserve significant geoarchaeological proxies.
Archaeological excavations at Huaca Cortada documented the occurrence of four El Niño events throughout the second half of the second millennium BCE. The identified El Niño proxies are laminated muddy layers deposited from runoff water. They contain thin laminas of white paint, which formed as the rain washed off the painted, plastered surfaces of the temple walls. These layers are sandwiched between pre- and post-event structures and were exposed to weathering for a short time, as inferred from their sedimentological features. By the
14C method, a date of 1600–1450 BCE is established for the earliest El Niño proxy of Huaca Cortada. The age of the subsequent events lies between 1100 and 900 BCE, as inferred from pottery finds. Finally, it must be noted that, as Caballo Muerto is surrounded by
quebradas, changes in the shape of the braided streams for each event may be supposed, which would be significant modifications for a desert environment. Such a question should be addressed with geological studies. Nevertheless, according to Nesbitt [
120], the communities of the Initial Period were able to rapidly reconstruct and enlarge buildings, and thus “
the social, religious and economic mechanisms that allowed for the mobilisation of labour […] were not negatively impacted by El Niño”.
The geoarchaeological record of the middle Moche Valley preserves a frequent recurrence of El Niño also for the two millennia of the Common Era. Huckleberry and Billman [
92] describe 13 ENSO-related flood and debris flow deposits beneath the present floodplain surface of Quebrada de los Chinos, which are younger than 2000 cal y BP. The findings of these authors seem to confirm the inference by Nials et al. [
15] on apparent fluvial landscape changes due to severe precipitations (see
Section 5.2), and reflect the late Holocene increases in El Niño activity [
10,
55]. With reference to methodological issues, the authors underline the need for a correct correlation between the geological characteristics of the deposits and the causative climatic events to make strong hypotheses [
92].
5.4. Casma and Cerro Sechin (Lower Casma Valley)
With reference to radiocarbon dating, Wells [
17,
82] analysed in detail the alluvial terraces system, which characterizes the lower Rio Casma Valley (
Figure 1). The system presents three floodplain surfaces whose ages of formation are between 3000 and 200 cal y BP (see Table 1 and Figure 4 of Reference [
17], pp. 1135–1136), thus suggesting significant geomorphological changes during the late pre-Hispanic time driven by river sediment transport and tectonic uplift. These results are consistent with data on the beach ridge accretion of the Pampas Las Salinas [
79,
80], about 50 km northwest of the Rio Casma mouth. From a geological point of view, such a landscape dynamics is not surprising for valleys and coasts of the Peruvian desert. Late Pleistocene and Lower Holocene were, for example, earlier periods throughout which El Niños severely impacted coastal Peru, leaving signatures on ancient landscapes (see, e.g., References [
19,
85,
93]).
Thirteen flood deposits younger than 3.2 ka are recorded along the stratigraphic section of Wells [
17,
82], six of which were deposited before the conquistadors first arrived in 1532. The upper two and the lower two of these latter are
14C dated by means of incorporated organic material. The author, however, does not recognize the ca. 1100 event (i.e., The Chimu Flood) defined by Nials et al. [
14] in the Moche Valley (see
Section 5.2). She focus mainly on the recurrence of the hydrogeomorphic events rather than their strength and suggests, for the considered time interval, the increase in frequency of flooding and the occurrence of an event “
much larger than that which occurred during 1982–1983” at least once every 1000 years. This hypothesis can be explained as follows: (a) the Mega El Niño events [
19,
80] actually exist and cause flood disasters in coastal Peru and possibly extreme climatic anomalies worldwide; or (b) the rainfall associated with the El Niño event is distributed such that extraordinary floods occur near Casma once every 1000 years [
17]. Clearly, this question is crucial for the topic treated in the review.
Complementary evidence on the paleoflood record of the Casma district can be provided by archaeological data from Cerro Sechin site (
Figure 1). However, only one laminated mud deposit ascribed to runoff water has been archaeologically dated (see Samaniego et al. [
78]) and may be correlated to a terraced alluvial deposit
14C dated at 1200 BCE by Wells [
82]. Likely, new geoarchaeological research in this area would help in providing significant data on landscape change due to El Niño events.
5.5. Old Temple of Pachacamac and Urpi Kocha Lagoon (Lower Lurin Valley)
According to Franco and Paredes, the Old Temple of Pachacamac (
Figure 4) was abandoned around 600 CE due to heavy rains that washed away and damaged the adobe walls of the building while runoff water deposited around thick layers of mud [
90]. These authors argue that an unusual climate event triggered temple modifications and led to the development of new social trends and the introduction of architectural elements from the Andes Highlands. Later, Franco stated “
there is no doubt that the rains that caused this event correspond to a Mega Niño” [
134], which is now referred to also as the 6th-century El Niño event. The sedimentological and palynological study by Winsborough et al. on the near Urpi Kocha Lagoon (located 0.4 km west of Pachacamac) confirms such a suggestion [
23]. These last authors found, in cores extracted from the bottom of the pool, evidence of three major floods associated with El Niño events in late pre-Hispanic time, the middle event dated between 436 and 651 CE (see Table 5 of Winsborough et al. [
23], p. 611).
The El Niño signature left at Pachacamac as recognized and interpreted by Franco and Paredes [
90] is similar to the ones described by Samaniego et al. [
78] and Wells [
82] at Cerro Sechin (see
Section 5.4) and by Nesbitt [
120] at Caballo Muerto (see
Section 5.3), respectively. It is apparent that, in late pre-Hispanic settlements, El Niño events may have left traces in the geoarchaeological record likely corresponding to landscape changes. Nevertheless, as carried out from further archaeological excavation, the adobe constructions of Pachacamac were seriously affected by repeated torrential rains up to the twentieth century, and especially by the 1925/1926 El Niño event [
135]. The erosive processes caused by these rains have partially erased the traces of the most ancient events, making the reconstruction of the paleoflood record at the lower Lurin Valley more difficult.
The validity of the hypothesis that one, or even two, El Niño events affected the central coastal Peru around 600 CE is confirmed by the findings of Mauricio [
136] at Huaca 20 site (Maranga archaeological complex, lower Rimac Valley, 30 km north-west of Pachacamac). Within such a site, now incorporated into the southern suburbs of Lima, the author describes two destructive floods separated by a phase of reconstruction during the Middle Horizon (
14C dated between 550 and 690 CE). Despite the limited size of the site, several geoarchaeological sections showing the above sequence are reported as basic data.
Other areas in the lower Lurin Valley offer opportunities for future research on the relationships between landscape modifications and El Niño events in the Initial Period and Early Horizon. Archaeological studies at the sites of Mina Perdida (5 km north-east of Pachacamac) and Manchay Bajo (10 km north-east of Pachacamac) have revealed the existence of debris flow, flood, and slack-water deposits in the respective geoarchaeological record [
137,
138]. A number of
14C dating on organic remains extracted from the stratigraphic successions provide chronological references for the El Niño signatures.
5.6. Pampa de Palpa (Valleys of Rio Ica and Rio Grande)
The
pampa southwest of Palpa (
Figure 1) constitutes the northern margin of the Atacama desert, and was inhabited by several cultures during the pre-Hispanic times. It has been the object of different studies on the development of loess deposits, paleosoils, and alluvial terraces that provide insights about the long-term relationship between climate processes and desert landscape changes [
4,
101,
106,
118]. Moreover, this hyper-arid region includes relevant sites of interest for the present review and is next to the ceremonial center of Cahuachi (which will be discussed separately in
Section 5.7).
Located at the lower valley of Rio Ica, the Samaca fluvial plain (
Figure 1) is currently a riparian oasis that hosts numerous archaeological and paleobotanic remains. The latter are mainly constituted by partially fossilized trunks of the phreatophyte
Prosopis, as established by Beresford-Jones et al. [
103,
104]. The geomorphological, archaeological, and paleoenvironmental evidence gathered by these authors allows them to state the occurrence of a major El Niño event “
at some time toward the end of the Early Intermediate Period, which spread a deep fluvial layer across the upper Samaca Basin, caused the river to cut some 5 m down into its floodplain and had catastrophic effects upon [a] canal system” [
103]. Such a result is consistent with the conclusions of other studies made in northern and central Peru [
23,
84,
86]. However, it explicates part of the above landscape modification, as well as of the decline in the settlement. In fact, according to [
103,
104,
109], human-induced gradual destruction of the
Prosopis forest in pre-Hispanic time would have considerably increased the exposure to the flood hazard of landscape and settlement, making them more vulnerable to severe events. The deforestation process would culminate during the Middle Horizon, causing the final abandonment of the settlement [
104].
Later, large urban settlements developed in the Late Intermediate Period (1000–1400 CE) within the
pampa of Palpa, the most prominent example of what is the so-called
Ciudad Perdida de Huayuri that, according to Hanzalova and Pavelka [
117], was destroyed by an El Niño event. However, no indications or evidence of such an episode were found in other articles identified by citation database queries (
Section 4,
Table 3 and
Table 4). It must be observed that Eitel and their colleagues [
4,
106] posit an abandonment of such a “lost city” as a consequence of the depletion of water reserve, likely related to a climate change. The above questionable information in Ref. [
117] may be due to a gap in the knowledge in the literature.
5.7. Cahuachi Ceremonial Center (Middle Nazca Valley)
According to Grodzicki [
21,
139], the archaeological site of Cahuachi (
Figure 5) was affected by three catastrophic floods caused by El Niño between 2100 and 1000 BP. These studies support the major landscape changes ascribed to El Niño events in late pre-Hispanic time for the whole coastal desert of Peru. The second event would have even caused the collapse of the Nasca Culture (around 600 CE), while the third completely buried the ceremonial center [
20,
21]. In support of their hypothesis, the author describes conglomerate deposits that would result from deposition of exceptionally large, fluid debris flows. Finally, the landscape of Cahuachi would have been re-shaped (and the monumental building unearthed) throughout the last millennium. A necessary assumption for this argument is that abandonment and burial of the ceremonial center preceded the construction of the Nazca Lines [
20]. However, such geoglyphs, created on the desert pavement by removing colored pebbles and leaving the underlying reg soil exposed, are mainly dated from 400 BCE to 600 CE [
140,
141,
142,
143].
Dealing with the second catastrophic flood inferred by Grodzicki [
139], Silverman [
6] observes that around 600 CE “
Cahuachi had ceased to function as the great early Nasca ceremonial center”. In their review on the occurrence of El Niño events, Machare and Ortlieb [
56] consider the hydrogeomorphic events posited in References [
20,
21] no more than possible local climate change indicators. Nevertheless, the main question on the reliability of the Grodzicki’s hypothesis comes from the research group headed by Eitel and Machtle [
4,
101,
106]. Starting from the observation of “
the good state of preservation of geoglyphs, even where they cross valleys or erosion rills”, such authors “
disagree with ideas that catastrophic El Niño events destroyed the Nasca Civilization” [
4]. Again, as a result of extensive field surveys, they remark how the integrity “
of the line exemplarily illustrates that the valley floor has never been flooded since the Nasca period. Only a small channel in the foreground provides evidence for weak episodic runoff events during the past two millennia” [
106]. Finally, these authors also extend their conclusion to the contiguous
pampa of Palpa (see
Section 5.6).
To further test the consistency of the hypothesis of Grodzicki, stratigraphic and petrographic analyses of upper bedrock and surficial cover of the ceremonial center were carried out by Delle Rose et al. [
26]. The results of this study show that the conglomerate deposits interpreted by Grodzicki as signatures of El Niño events, belonging instead to the Tertiary–Quaternary clastic succession that forms the regional substratum. Moreover, such coarse-size sediments are a source of pebbles and cobbles, which form the desert pavement. The inconsistency of Grodzicki’s hypothesis was likely due to knowledge gap in geological stratigraphy [
26].
Recent archaeological excavations on
Templo Sur have shed new light on hydrogeomorphic events that damaged the adobe brick buildings at Cahuachi during the development of the Nasca Culture. In fact, the inferences of Orefici [
144] about two torrential rains that partially destroyed roof and walls of the temple next to the end of the fourth century CE, a time with no El Niño events reported in the reviewed literature, lead us to reconsider both the climatic pattern responsible for the events and the type of Earth surface response expected for the coastal desert. It must be noted that samples gathered at Cantalloc (20 km east of Cahuachi) by Pavelka et al. [
122], considered by these authors remains of an ancient flood, are
14C dated between 47 and 480 CE.
5.8. Moquegua Valley and Quebrada Miraflores
Geoarchaeological studies throughout Moquegua Valley and surrounding areas [
22,
24,
51,
84] were mainly aimed to argue the relationships between El Niño events and human responses rather than possible landscape processes [
91]. Nevertheless, data exposed in the reviewed works allow one to argue for abrupt geomorphological changes to the desert environment. As a matter of fact, already thirty years ago, Moseley et al. [
84] stated that “
serious landscape modification should indicate extreme ENSO conditions or ancient ‘Mega-Niño’ phenomena”.
Three main depositional units confidently ascribed to El Niño events have been identified and dated by the authors. Two of them have a late pre-Hispanic age, while the third (named Chuza Unit) is dated at 1607–1608 CE [
96]. The oldest
14C dated unit (730–690 CE) is formed by debris flow deposits (identified within the dry Rio Muerto sub-basin [
24] (see
Figure 6)), and flood deposits recognized elsewhere [
51]. As it signaled fast and extensive regolith mobilization, this unit constitutes evidence of a singular hydrogeomorphic event. For what concerns the regolith production, in a highly seismic region such as coastal Peru, it is increased by the shattering of the landscape during the frequent earthquakes that produce pervasive ground cracking and microfracturing of hill-slope materials [
84,
96].
The second depositional unit of interest for this review is a horizon marker referred to as the Miraflores Flood event [
84]. It is formed by debris flow deposits and flood deposits indirectly dated between 1350 and 1370 CE by ice core data of the Quelccaya Ice Cap [
22,
145,
146]. Such an age is confirmed by radiocarbon dates of samples taken at Rio Muerto [
51] (
Figure 6). At Quebrada Miraflores (
Figure 7), the thickness of the marker reaches 1.2 m, which is a higher order of magnitude compared with any other flood deposits preserved in the geological record, including the Chuza Unit.
Figure 7.
Quebrada Miraflores; 2021 satellite image (17
25′49″–17
26′36″ S, 71
22′22″–71
21′42″ W). The extent of the Miraflores Unit according to Satterlee et al. [
22] is marked by a red line and the location of Chiribaya settlement by a blue circle (see text).
Figure 7.
Quebrada Miraflores; 2021 satellite image (17
25′49″–17
26′36″ S, 71
22′22″–71
21′42″ W). The extent of the Miraflores Unit according to Satterlee et al. [
22] is marked by a red line and the location of Chiribaya settlement by a blue circle (see text).
In this last area, mobilized sediments spread laterally out of the braided streams and up the ravine walls before descending to the sea across the coastal plain. Large clasts and boulders with a diameter of up to 3 m were also moved across the coastal plain by the Miraflores Flood [
22]. Such a hydrogeomorphic event would also be implicated in the collapse or decline of the Chiribaya culture [
111,
147]. The imprint on the Peruvian desert surface of this 14th-century El Niño event is confirmed by zooarchaeological research. Rubinatto et al. posit “
that the abundance of anuran remains” they found in Rio Muerto area may be related to the Miraflores Flood, since “
this event generated increased rainfall in the desert, creating conditions favorable for frogs and toads” [
130]. However, for a critical examination of such a hypothesis, these authors exhort further paleoenvironmental and zooarchaeological studies. In any case, the potential of Miraflores Unit as a horizon marker should be fully exploited in landscape reconstruction.
5.9. Discussion Summary
Landscape processes confidently ascribed to El Niño events are chronologically summarized in
Table 5. They are considered responsible for changes in late pre-Hispanic coastal Peru to varying degrees.
The above chronology may be tentatively compared with paleo El Niño proxies. By analyzing a high resolution marine sediment record about 60 km west of Lima, Rein et al. [
148] argued maximum of El Niño activity during the third and second millennium BP, while El Niño events would have been persistently weak during the Medieval Climate Anomaly (MCA, ca. 900–1300 CE). Yan et al. [
149], computing SOI (
Section 2) from different precipitation proxy records in Pacific Ocean, found negative index during MCA, which indicates general El Niño-dominated conditions. However, such differences in climatic reconstructions may reflect the geographic complexities of the Pacific coastal regions [
45]. As a matter of fact, SOI calculated for the Galapagos Islands (850 km off the coast of Ecuador) shows positive values during MCA in contrast with the ones of the western Pacific [
149]. The oxygen isotopes ratio (
18O) of ice core record from Quelccaya ice cap is depleted in
18O between 1100 and 1300 CE, thus highlighting a warming trend over the Andes [
150]. Such a finding is consistent with low El Niño activity during MCA [
148,
151]. Consequently, the hydrogeomorphic events that can be attributed to El Niño with the lowest uncertainties should be those that precede or follow the MCA.
As a result of the WoS and Scopus data mining and the subsequent full-text analysis, the first evaluable hydrogeomorphic signature is located at Caballo Muerto and
14C-dated at about 3.5 ka BP [
120]. However, it is in advance of the beginning of ENSO’s modern periodicity (
Section 2). Three more recent deposits dated around 3.0 ka and ascribed to El Niño events were also reported by this author (
Table 5). To date, the magnitude of landscape response to these events along the middle Moche Valley is not yet investigated, while the only identified surface process is mud deposition from runoff over archaeological structures (
Section 5.3). Moreover, the lower Lurin Valley also preserves geoarchaeological proxies [
137,
138] that can help shed light on the landscape modifications that occurred around the transition from the second to the first millennia BCE.
Around 1200 BCE, Casma Valley was affected by torrential rains, and thus, related proxies are preserved in both alluvial terrace sequences [
82] and archaeological records of Cerro Sechin [
78]. The hypothesis that both signatures were produced by the same El Niño event should be verified by further research (
Section 5.4). According to Wells [
17], the alluvial terrace system of the Casma Valley contains 13 flood deposits ascribable to late Holocene El Niños. Four
14C-dated events occurred in pre-Hispanic time, but unfortunately, no date fits well with events identified in other basins or sites (
Table 5). In contrast, one or more events transpiring about 600 CE have hit the valleys of Rio Moche (Huacas del Sol), Rio Lurin (Pachacamac), and Rio Ica (Samaca) (
Section 5.2,
Section 5.5 and
Section 5.6). Such evidence has led some authors to speak of a 6th-century Mega El Niño (see, e.g., [
21,
23,
134]).
Nevertheless, the recognition of one or more events to the transition from the first to the second millennium CE at some sites of north-central coastal Peru (
Table 5) has strengthened the paradigm of the 11th-century Mega El Niño (see, e.g., [
21,
56,
152]). The seminal reconstruction of the Chimu Flood by Nials et al. [
14,
15] has greatly influenced the literature, but it should be tested by further research, and the related deposits should be absolute-dated. Furthermore, the age of the ethno-historical Naymlap Flood remains unclear [
84,
132]. It ranges from around 1000 CE (see, e.g., [
16,
23]) to around 1350 CE (see, e.g., [
17,
153,
154]). As a matter of fact, Naymlap Flood “
may refer to different floods in different valleys at different times during the Late Intermediate Period” [
7].
If on the one hand, El Niño events had catastrophic consequences associated with floodplain stripping, irrigation system damage, and settlement destruction, on the other hand, they also provided cultural opportunities, triggered technological innovation, strengthened community resilience against hydrogeomorphological events, and played important roles in replenishing and maintaining groundwater resources throughout the coastal desert [
51,
88,
138]. Nevertheless, extensive human intervention (desert reclamation,
Prosopis deforestation) has likely caused some variations in landscape response, also increasing exposure to extreme events of settlements and infrastructures (
Section 5.2 and
Section 5.6).
Authors have deeply investigated the landscape response and destructive effects of the Miraflores Flood, the last hydrogeomorphic event ascribed to a “Mega” El Niño in pre-Hispanic times [
22,
24,
51,
84,
111,
147] (
Section 5.8). With the arrival of Spanish conquistadors, the production of historical documents begins, so the sources of knowledge on El Niño events are enriched with a fundamental element [
34,
56,
81,
155]. However, it is only with the systematic observations of the warm ocean current running along the central Pacific coast of South America (El Niño in its initial meaning, see
Section 2) that all the necessary criteria to assess El Niño events are available.
A final consideration must be made on the limitation of the above finding due to the use of WoS and Scopus. Papers published in journals not indexed in these citation databases are obviously excluded. Moreover, as remarked upon in
Section 3, documents written in English are favored over those written in other languages in data mining from WoS and Scopus [
68]. Such problems may have partially affected this review. However, in the discussion on the main study areas (
Section 5),documents written in Spanish and published in (not WoS- and/or Scopus-indexed) Peruvian journals and books ([
131,
134,
135,
136]) were also used, to the best of the author’s knowledge. Clearly, the result may be hereafter improved and updated. On the other hand, data mining from other databases, systematic reading of papers written in Spanish and published in Peruvian journals, and the use of web search engines such as Google Scholar could be goals of future investigations.