Restoration of weed communities in abandoned rice paddy fields in the Tama Hills, central Japan Susumu Yamada, Satoru Okubo, Yoshiko Kitagawa and Kazuhiko Takeuchi Department of Ecosystem Studies, Graduate School of Agricultural and Life Sciences, The University of Tokyo, Yayoi 1-1-1, Bunkyo-ku, Tokyo 113-8657, Japan Name and address for correspondence: Susumu Yamada Department of Ecosystem Studies, Graduate School of Agricultural and Life Sciences, The University of Tokyo, Yayoi 1-1-1, Bunkyo-ku, Tokyo 113-8657, Japan Tel.: +81-3-5841-5052, Fax: +81-3-5841-5072 E-mail address: [email protected] Abstract Since paddy weed communities are declining with intensification and abandonment of agricultural activities, protecting these species has become important. The restoration of cultural ecosystems normally includes the concomitant recovery of indigenous management practices. Given that the abandonment of paddy fields is due largely to cost and the shortage of labour, reintroduction of annual farming is unlikely even if paddy fields are restored. Hence, to confirm the occurrence of species assemblages typically represented in rice paddy 1 fields (typical paddy weeds: TPWs) and to determine how to decrease the expense and management burden, we instituted a restoration program in which traditional rice culture was restarted with some modifications such as introducing fallow periods in paddy fields abandoned for more than 10 years. We also investigated cultivated rice paddy fields where local farmers practise low-intensity farming and newly abandoned paddy fields as references. By comparing the floristic composition between restored and reference fields, we sought to meet two objectives: (1) to evaluate the efficacy of the restoration program in terms of the species composition of paddy weed communities in restored paddy fields, and (2) to identify the optimum fallow period to decrease the expense and management burden. Restarting agricultural practices in abandoned paddy fields successfully encouraged TPWs. Moreover, restoration of the TPW community was rapid, with no resilient or alternative state in the restoration program. Soil tillage and water management appeared to be major factors determining the success of the restoration. Tillage restricted the dominance of rhizomatous perennials and provided habitats for low-stature summer annuals. Reintroduction of submergence as a water management practice generally inhibits the occurrence of most TPWs. However, the unlevelled soil surface in the present study provided suitable conditions for these species. One-year fallow produced the highest abundance of TPWs in a transient stage influenced by antecedent effects of rice culture such as drainage after rice harvest, delay of competitive species incidence, and lack of submergence as an agricultural treatment. The number of TPWs declined steadily with increasing fallow period. This evidence indicates that 2 to maintain paddy weeds in the emergent flora, it is appropriate to create 1-year fallow conditions; that is, to cultivate every other year. Key words Agricultural practice; Cultivation; Fallow; Mowing; Wetlands 1. Introduction Rice (Oryza sativa L.) has been cultivated for a long time in Asia. In Japan, rice paddies are the most common production farmlands, covering about 55% of the total agricultural land in 2000 (MAFF, 2003), and are thus the main component of agricultural landscapes. Although the water level is controlled artificially, rice paddies have recently been evaluated as wetland habitat for hygrophytes and hydrophytes known as ‘paddy weeds’ (Sekioka et al., 2000) and for water birds (Lane and Fujioka, 1998) and frogs (Osawa and Katsuno, 2001). However, they have been altered drastically in recent decades by socio-economic changes. As in agricultural landscapes in Europe, biodiversity is declining with intensification and marginalization of agricultural activities (Robinson and Sutherland, 2002; Hyvönen and Salonen, 2002). To enhance rice productivity and farmers’ work efficiency, many paddy fields, especially those located in wide alluvial lowlands, have been equipped with modern drainage systems 3 and made drier during the non-cultivation season (Hasegawa and Tabuchi, 1995). This change in water condition has decreased the value of the paddy fields’ ecological function as wetland habitat, encouraging less hydrophilic vegetation and reducing floristic species richness (Arita and Kobayashi, 2000). Meanwhile, as a result of shifting economics and the decline of the rural population since the 1970s, a substantial number of paddy fields, especially those located in mountainous and hilly areas, are no longer cultivated (MAFF, 2001). Many studies reported the establishment of wet grassland characterized by Phragmites australis and Isachne globosa (Hakoyama et al., 1977; Matsumura et al., 1988; Ohkuro et al., 1996; Comín et al., 2001) or wet woodland characterized by Salix koriyanagi and Alnus japonica (Hayakawa and Takahata, 1975; Shimoda, 1996; Lee et al., 2002) in the process of secondary succession when land management was not applied and if fields remained wet. Of course, secondary succession is a complex multifactorial process. The process depends on light conditions (Kang et al., 2004), seed sources from surrounding vegetation (Shimoda, 1996), climate (Lee et al., 2002), and soil moisture (Hakoyama et al., 1977), so correspondingly there are other dominant species. Nevertheless, species representative of cultivated paddy fields tended to decrease in most of these processes, even if the fields remained wet (Matsumura et al., 1988; Shimoda, 1996). With these situations, many paddy weed species have become threatened. Kasahara (1951) defined 191 ‘paddy weeds’ that needed to be controlled in those days in Japan. Of those, 110 species were especially problematic because they are hard to control in rice culture. 4 Yet now, 10 of those ‘problematic’ weeds are listed in the Japanese Red Data Book (Environment Agency of Japan, 2000) as threatened, and even more species are listed at a regional level. Therefore, immediate actions should be launched to preserve paddy fields in a wet condition and to establish alternative management schemes for restoring the specific conditions in which paddy weeds grow. Successful restoration faces constraints. Recent reports suggest that some degraded systems shift to a new state (a resilient or alternative state) that cannot be restored to the previous condition or disturbance regime (Kleijn, 2003; Suding et al., 2004). Strong feedback between biotic factors and the physical environment can alter the efficiency of management efforts because of constraints such as eutrophication and acidification (Pywell et al., 1995; Willems, 2001; Walker et al., 2004), landscape connectivity (Hutchings and Booth, 1996; Bischoff, 2005), loss of species pool (Bakker and Berendse, 1999), and shifts in species dominance (Andersen et al., 2000). Another problem is more practical. The restoration of cultural ecosystems normally includes the concomitant recovery of indigenous management practices (Society for Ecological Restoration International Science & Policy Working Group, 2004). Given that the abandonment of paddy fields is due largely to cost and the shortage of labour, annual farming is unlikely even if paddy fields are restored. In Japan, several restoration programs for conserving rare arable weeds have been carried out by restarting cultural practices in abandoned paddy fields (Asami et al., 2001; Shimoda and Nakamoto, 2003). However, it is still unclear whether vegetation communities 5 represented in rice culture are totally restored by agricultural treatments in formerly abandoned paddy fields. Furthermore, few studies have assessed the effectiveness of management with regard to labour and cost efficiencies. A potentially suitable alternative is cultivation with fallow periods (i.e., producing crops only every few years). Species richness is likely to be low in cultivated paddy fields because most weeds are eradicated during cultivation; richness would be higher in fallow fields, as in the case of upland fields (Degn, 2001; Harmer et al., 2001). During long fallows, however, it might decline. Confirming the patterns and pace of vegetation shifts will allow us to establish effective management schemes. In the present study, the restoration process restarted traditional rice culture with some modifications such as introducing a fallow period in paddy fields abandoned for more than 10 years. We monitored species assemblages typically represented in rice paddy fields in hilly alluvial lowland to evaluate the restoration effort. We also investigated cultivated rice paddy fields where local farmers practise low-intensity farming, and newly abandoned paddy fields as references. By comparing the floristic composition between restored and reference fields, we sought to meet two objectives: (1) to evaluate the efficacy of the restoration program in terms of the species composition of paddy weed communities in restored paddy fields, and (2) to identify the optimum fallow period to decrease the expense and management burden. 2. Materials and methods 6 2.1. Study area The study area is located in the Tama Hills, 30 km west of Tokyo, at latitude 35°35′N, longitude 139°25′E. The mean annual precipitation at the nearby Hachioji meteorological station is approximately 1540 mm. The mean annual temperature is 14.3 °C, the mean minimum is 3.2 °C (January), and the mean maximum is 26.1 °C (August). The bedrock of the Tama Hills consists of semi-consolidated or unconsolidated sedimentary rocks from the Pliocene to Middle Pleistocene period, commonly known as the Kazusa Group. This group comprises several formations of fluvial gravels, tidal flats and inner bay silts, and littoral to upper neritic sand from the lower to the upper part (Takano, 1994). The layer underlying the valley bottoms is an impermeable mudstone. Most soils in the area contain tephra; the soil type of the paddy fields on the valley bottoms is Andic Gleysols. Hill ridges are about 30 m higher than the adjacent valley bottom. Hill slopes are mainly dominated by deciduous, broad-leaved forests, characterised by Quercus serrata, Q. acutissima and Castanea crenata, which were formerly coppiced (Okubo et al., 2005). Rice is cultivated mostly in valley bottoms (Fig. 1). No intensive land improvement (land consolidation and development of water control system) is practised because of the low economic efficiency in such small valleys; hence, paddy fields stay irregular in shape and drainage remains poor in all seasons. Although a submerged condition is necessary during the cultivation period, drainage is needed in order to make the process of rice cultivation more efficient. The fields are surrounded by embankments approximately 50 cm wide, and are 7 aligned from valley head to mouth on a terraced slope of 1 to 4%. The size of each parcel is very small, ranging from 180 to 400 m2, and averaging 300 m2. The fields are irrigated with natural water resources such as spring water from the valley head and rain water. 2.2. Restoration 2.2.1. Restoration of abandoned paddy fields The restoration site lies in a small alluvial valley (Fig. 1), inside the 33-ha Zushi-Onoji Historic Environmental Conservation Area, which was proclaimed a Greenery Designated Conservation Area by the Tokyo Metropolis in 1978. According to aerial photo interpretation and interviews with local farmers, these paddy fields were partially abandoned in the 1970s and completely abandoned by the middle of the 1980s. Before the restoration, abandoned paddy fields were dominated by P. australis, I. globosa, and Leersia sayanuka, with scattered Alnus hirsuta var. sibirica and Salix spp. (Kitagawa, 2003). Footpaths, embankments, and canals around the paddy fields had collapsed. Restoration practices started in the late summer of 1996. The Tama Environment Office of the Bureau of Environment, Tokyo Metropolis, and a local management organization formulated the management design. Since the management organization consists mostly of local farmers who used to own and cultivate the fields, their indigenous knowledge was drawn in the restoration to recover previous conditions and in the subsequent management. At the beginning of the restoration, all tall vegetation was cut and burnt in an area of 0.4 ha. 8 Collapsed footpaths and embankments were mounded as high as they had been before abandonment, and sedimented canals around the paddy fields were re-excavated to re-establish water management. By 1997, only 4 neighbouring paddy rice parcels in the lower part of the valley had been completely restored by restarting paddy cultivation (Figs 1 and 2). A lack of human resources and budget limited the restoration, although the Tama Environmental Office and the management organization recognize that all parcels should be restored to preserve the traditional rural landscape (satoyama landscape; Fukamachi et al., 2001). 2.2.2. Treatment of restored paddy fields All 4 restored parcels were managed under different time-series treatments from 1997 until 2003, when our monitoring was terminated. There were 2 types of treatment combination, which differed mainly in the paddy cultivation phase: cultivation with or without rice. Three of the 4 restored parcels were managed with the treatment combination of paddy cultivation with rice and fallow (cessation of cultivation), henceforth referred to as 'restored parcels with rice culture'. For the rice culture phase, the management regime was almost the same as that used before rice culture was abandoned. Fields are ploughed in May. A few days before transplanting in June, they are ploughed again, flooded to about 15 cm deep, puddled, and levelled (‘surface soil puddling’). Then rice is transplanted. The water level is maintained for several weeks, and weeds are manually eradicated twice. At the end of July, the fields are drained via ditches dug when the rice starts maturing. Finally, rice is harvested in October. 9 Thereafter, the parcels are drained, but the surface soil remains wet. The water level gradually becomes higher owing to a constant supply of water from springs, and by the next summer the surface is naturally submerged by up to 5 cm. The difference from the traditional rice culture was that soil tillage after rice harvest was not carried out, because of the shortage of labour and funds. The fallow treatment was applied after rice was grown at least once. In this treatment, water level was not controlled artificially, so parcels were submerged about 5 to 10 cm deep during the treatment. During this phase, the parcel was mown with a shoulder-held mower twice a year (end of May to middle of June, and again in late August) to inhibit the dominance of macrophytes and woody species. Two of the 3 parcels had the same treatment combination. Rice culture was performed from 1997 to 2000 and in 2003, with a break from 2001 to 2002. In the 3rd parcel, rice culture was performed in 1997 and 1999, with breaks in 1998 and from 2000 to 2002. In 2003, part of the parcel was cultivated. In the 4th restored parcel, paddy cultivation without rice was performed in 1997 ('restored parcel without rice culture'). The treatment was the same as that with rice except for the rice and the water management: water was not drained, but remained about 15 cm deep throughout the treatment. After the treatment, the parcel was left to lie fallow and untouched for 5 years, from 1998 until 2002. In 2003, the vegetation was cut. 2.2.3. Monitoring in restored paddy fields To accomplish our two objectives, we monitored chronosequential changes of vegetation in 10